Engineering Design Foundations and Requirements

Introduction

Machine component design is the disciplined process of converting a need into a safe, functional, manufacturable, economical, and maintainable physical product. It connects mechanics, materials, manufacturing, economics, safety, human factors, standards, and project management. A machine component may be a simple shaft, key, bearing housing, fastener, spring, gear, or welded bracket; however, its design cannot be judged only by whether it can carry a calculated load. It must also perform its intended function throughout its service life and within the constraints imposed by users, production systems, regulations, and the environment.

Design is therefore more than sizing. Sizing determines dimensions after the design problem has been defined. Design definition asks what the component must do, how it interacts with neighboring parts, what can go wrong, how it will be produced, and how its performance will be verified.

This chapter establishes the vocabulary and reasoning methods used throughout machine component design. It introduces requirements, functional decomposition, constraints, load paths, standards, codes, safety factors, design reviews, and the iterative transition from concept to production.

Learning Objectives

After studying this chapter, you should be able to:

1. Distinguish design, analysis, verification, validation, and production engineering.
2. Translate a user need into measurable design specifications.
3. Decompose a machine into functions, subfunctions, interfaces, and components.
4. Identify load paths, constraints, failure modes, and relevant standards.
5. Explain the purpose and limitations of factors of safety.
6. Describe a staged, iterative design process from concept to production.
7. Participate effectively in design reviews and communicate engineering decisions.

1. Design Terminology and the Nature of the Design Problem

A machine is an assembly of interacting parts that transmits or transforms energy to perform useful work. A component is a relatively distinct part or subassembly of a machine. A system is a set of components and interfaces organized to achieve an overall purpose. An assembly is a collection of parts joined for operation, service, or manufacture.

A design requirement is a condition that the product must satisfy. Requirements may concern performance, dimensions, loads, temperature, life, safety, cost, noise, maintainability, or legal compliance. A constraint limits the possible solutions. For example, a requirement may specify a torque of 500 N·m, whereas constraints may limit the shaft diameter, available material, manufacturing process, or mass.

A design specification converts a general need into measurable targets and acceptance criteria. “The coupling must be reliable” is not an adequate specification. “The coupling shall transmit 500 N·m continuously, 750 N·m for 10 s, at 1,800 rpm, for 10,000 operating hours, with no hazardous fragment release” is much more useful because it can guide calculations and tests.

Several terms must be separated carefully:

• Conceptual design establishes functions, architecture, and alternative principles.
• Embodiment design establishes arrangement, interfaces, materials, dimensions, and manufacturing concepts.
• Detail design defines complete geometry, tolerances, surface specifications, drawings, bills of materials, and inspection requirements.
• Analysis predicts behavior using models, calculations, simulation, or testing.
• Verification asks whether the design output satisfies specified requirements: “Did we build the product right?”
• Validation asks whether the product satisfies the actual user need in its intended context: “Did we build the right product?”
• Manufacturability is the degree to which a design can be produced consistently, economically, and with available processes.
• Reliability is the probability that an item performs its required function for a specified time under stated conditions.
• Maintainability concerns the ease and time required to inspect, repair, replace, or restore a component.

Design problems are usually underdetermined: many geometries can meet the same functional requirement. They are also coupled: changing a shaft diameter can affect bearings, seals, housing size, mass, cost, and assembly. Consequently, a design is not a single calculation but a sequence of decisions under uncertainty.

2. From Need to Design Requirements

The design process begins with a need, opportunity, defect, or safety concern. A need is expressed in the language of users or organizations; engineering requirements express it in measurable technical terms. The transformation should preserve the intended purpose while removing ambiguity.

A useful requirement should be:

• Necessary: directly related to the product purpose.
• Unambiguous: interpreted in only one reasonable way.
• Measurable or testable: associated with a method of verification.
• Traceable: linked to a source, such as a customer, regulation, hazard analysis, or interface.
• Realistic: achievable with available technology and resources.
• Prioritized: classified as mandatory, desirable, or negotiable.

Requirements are commonly organized into categories:

A requirements traceability matrix links each requirement to its source, design feature, analysis, test, and final evidence. Traceability prevents a common failure in which a design is optimized for easily calculated quantities while neglecting difficult but essential requirements such as access for maintenance or safe behavior after loss of power.

Requirement conflicts and prioritization

Requirements often conflict. A lighter shaft may reduce material cost but increase deflection. A tighter fit may improve accuracy but complicate assembly. Greater enclosure strength may improve safety but increase mass. The designer should expose these conflicts rather than silently resolving them.

One practical method is to classify requirements as:

1. Non-negotiable: legal, safety, or essential interface conditions.
2. Target: preferred values used for optimization.
3. Limit: upper or lower bounds that cannot be exceeded.
4. Assumption: provisional information requiring confirmation.

Assumptions must be recorded because an apparently successful design may depend on an unverified duty cycle, material property, operator behavior, or environmental condition.

3. Functional Decomposition and System Architecture

Functional decomposition breaks an overall purpose into smaller functions. For a powered conveyor, the top-level function may be “move product at controlled speed.” Subfunctions may include:

1. receive electrical energy;
2. convert electrical energy to rotation;
3. transmit torque;
4. support rotating members;
5. reduce or increase speed;
6. engage the load;
7. guide the product;
8. protect users;
9. dissipate heat;
10. permit inspection and maintenance.

Each function can be assigned to one or more physical components. The mapping is not always one-to-one. A housing may support a bearing, contain lubricant, locate a shaft, and provide guarding. Conversely, a function such as “prevent excessive speed” may require a sensor, controller, brake, and software.

Functional decomposition is valuable because it prevents premature fixation on a familiar component. Instead of beginning with “select a larger shaft,” the designer asks what function is failing: torque transmission, alignment, stiffness, fatigue resistance, or attachment. This distinction often reveals a better solution, such as changing the load path, adding a support, or reducing shock loading.

A functional architecture describes how functions interact. A physical architecture describes the components and interfaces that realize those functions. Important interfaces include:

• mechanical interfaces: fits, fasteners, splines, keys, bearings;
• energy interfaces: electrical, hydraulic, pneumatic, thermal, or mechanical power;
• information interfaces: sensors, control signals, feedback;
• environmental interfaces: seals, vents, cooling, contamination barriers;
• human interfaces: access, controls, visibility, ergonomics.

Interface definition is central to successful design. Many field failures occur not because an individual component is weak, but because components are misaligned, incompatible, difficult to assemble, or exposed to loads that were omitted from the original analysis.

4. Load Paths and Free-Body Reasoning

A load path is the route through which force, torque, energy, or heat travels from its source to its destination. For a belt-driven shaft, torque may pass from the motor through a coupling, shaft, key or spline, pulley, belt, driven pulley, and supported shaft into the machine frame. Every element in that path must be considered.

Load-path analysis begins with a system boundary and a free-body diagram. The designer identifies applied loads, reactions, inertia forces, friction, thermal effects, and external constraints. Loads should be classified by type and time behavior:

• static or slowly varying;
• repeated or cyclic;
• impact or shock;
• transient during start-up or stopping;
• random or vibration-induced;
• thermal or residual;
• environmental, such as wind, pressure, or contamination.

For a shaft transmitting torque T at rotational speed n, ideal power is:

Here, P is power in watts, T is torque in newton-metres, ω is angular speed in radians per second, and n is rotational speed in revolutions per second. If speed is given in revolutions per minute, then:

where N is rotational speed in revolutions per minute. These relations assume ideal power transmission; real systems require efficiency η, giving P_out = ηP_in.

For a simply supported beam with a central point load W and span L, the maximum bending moment is:

This is a special case, applicable only to a statically determinate, simply supported beam with one central concentrated load and negligible support settlement. It should not be used for a cantilever, distributed load, or multiple loads without modification.

For a circular solid shaft of diameter d subjected to a torque T, the maximum elastic torsional shear stress is:

Here τ_max is shear stress in pascals, T is torque in newton-metres, and d is diameter in metres. The equation assumes a homogeneous, isotropic, circular shaft in elastic torsion, with Saint-Venant behavior and no stress concentration. Keyways, shoulders, corrosion, and manufacturing defects require additional consideration.

A load path is incomplete unless reactions and interfaces are included. For example, a bearing does not merely “support a shaft”; it receives radial and axial reactions, transmits them to a housing, and ultimately transfers them to the frame and foundation. The housing bolts and foundation must therefore be checked as part of the same path.

5. Constraints, Failure Modes, and Design Specifications

Constraints define the feasible design region. Common constraints include maximum envelope, minimum clearance, available tools, material availability, assembly direction, noise limits, temperature, budget, and delivery schedule. A constraint may be quantitative, such as a maximum mass of 25 kg, or qualitative, such as compatibility with existing maintenance practices.

Failure-mode thinking should begin early. A failure mode is the manner in which a component can fail, such as yielding, fracture, fatigue cracking, buckling, wear, seizure, leakage, loosening, overheating, corrosion, or excessive deflection. A component can remain unbroken yet fail functionally if it becomes too flexible, noisy, inaccurate, or difficult to maintain.

A simplified failure analysis asks:

1. What function can be lost?
2. What physical mechanism causes the loss?
3. What is the consequence?
4. How can the cause be prevented or detected?
5. What evidence demonstrates adequate control?

Methods such as failure mode and effects analysis (FMEA), fault-tree analysis, hazard analysis, and design-for-safety review formalize this reasoning. They do not replace engineering judgment; they make assumptions and responsibilities more visible.

A design specification should include operating cases. For a shaft, these may include normal torque, start-up torque, emergency stopping, misalignment, belt tension, thermal expansion, and accidental blockage. Design against only the nominal case can produce an unsafe result.

6. Standards, Codes, and Standardization

Standards are agreed technical documents that define terminology, dimensions, test methods, materials, tolerances, or procedures. Codes generally contain enforceable or quasi-regulatory requirements for safety and construction. Regulations are legally binding requirements established by authorities. Their exact status depends on jurisdiction and application; a standard is not automatically law, but it may become mandatory through legislation, contract, or referenced design rules.

Examples of standardization in machine design include preferred fastener sizes, bearing dimensions, fits, keys, threads, material grades, symbols, drawing practices, and testing procedures. Standard components reduce cost, shorten procurement time, simplify replacement, and improve interoperability. They also provide known performance data and established inspection methods.

Standardization should not be applied mechanically. A catalog bearing may be dimensionally convenient but unsuitable for contamination, speed, temperature, or life. A standard material grade may be unavailable in the required heat treatment. The designer must verify the edition, scope, units, assumptions, and applicability of every referenced document.

Good practice includes:

• identifying applicable standards during requirements definition;
• recording exact document titles and editions;
• distinguishing mandatory clauses from guidance;
• checking regional and industry-specific requirements;
• using standards for interfaces and safety-critical features;
• obtaining supplier certification and material traceability where necessary.

7. Safety Factors and Uncertainty

A factor of safety compares a capacity or limiting condition with a design demand. In its simplest form:

where n is the factor of safety, S_limit is the relevant limiting strength or capacity, and S_demand is the applied stress or load. If a yield-based criterion is used, a nominal allowable stress may be written as:

where S_y is yield strength and σ_allow is allowable stress. This formulation assumes that the chosen failure criterion is appropriate and that stress, material, geometry, loading, and environment have been represented adequately.

A factor of safety is not a universal measure of reliability. It may compensate for uncertainty in material properties, loading, manufacturing, modeling, or consequences of failure, but it does not correct a wrong load path or an omitted failure mode. Fatigue, fracture, buckling, wear, and thermal failure require their own models and criteria.

For a simple tensile member carrying a load F, with cross-sectional area A, the nominal normal stress is:

If F = 20 kN and A = 400 mm², then σ = 20,000/400 = 50 N/mm² = 50 MPa. If the material yield strength is 250 MPa and a nominal factor of safety of 2 is selected, the allowable stress is 125 MPa, so the simple yield check passes. This conclusion is limited: it does not address stress concentration, fatigue, connection failure, corrosion, buckling, or deflection.

Selection of n should reflect uncertainty and consequence. Higher values may be appropriate for poorly known loads or severe consequences, but excessive values can create unnecessary mass, cost, and inflexibility. Modern practice increasingly combines explicit load distributions, reliability methods, testing, and standards-based design rather than relying on one arbitrarily large number.

8. The Iterative Design Process

A robust design process is iterative rather than strictly linear:

1. Identify the need and stakeholders.
2. Establish requirements, constraints, and operating scenarios.
3. Decompose functions and define interfaces.
4. Generate several concepts.
5. Screen concepts for safety, feasibility, and value.
6. Develop the selected architecture.
7. Analyze loads, stresses, motion, thermal behavior, life, and risks.
8. Select materials and manufacturing processes.
9. Produce prototypes or representative tests.
10. Verify and validate.
11. Release drawings, specifications, and production documentation.
12. Monitor production and field performance, then improve the design.

The process contains feedback loops. Analysis may show that a concept cannot meet stiffness requirements. A prototype may reveal assembly difficulty. Production may reveal excessive variation. Field data may identify an unanticipated duty cycle. Each result should update requirements, assumptions, models, or geometry as appropriate.

Concept selection can use a weighted decision matrix, provided the criteria and weighting are justified. Typical criteria include safety, function, manufacturability, cost, mass, maintenance, schedule, and environmental impact. A numerical score does not make a decision objective by itself; it makes the basis for discussion explicit.

Design reviews

A design review is a structured evaluation by people with appropriate technical, manufacturing, quality, safety, service, and user knowledge. Effective reviews examine evidence rather than merely asking whether the design “looks good.” Typical review questions include:

• Are all requirements traceable to verification evidence?
• Are load cases and environmental conditions complete?
• Have interfaces and assembly sequences been checked?
• Are failure modes controlled or detectable?
• Are tolerances, materials, and finishes realistic?
• Can the component be inspected and maintained safely?
• Are changes controlled and documented?

Review gates may be called concept review, preliminary design review, critical design review, production-readiness review, and post-launch review. A review should produce decisions, action owners, due dates, unresolved risks, and approval status.

9. Transition from Concept to Production

The transition to production is an engineering activity, not an administrative handoff. A production-ready definition normally includes:

• fully dimensioned and toleranced drawings;
• material and heat-treatment requirements;
• surface finish and coating specifications;
• joining and assembly instructions;
• inspection and test plans;
• bills of materials and approved suppliers;
• packaging, storage, and preservation requirements;
• service and maintenance information;
• configuration and change-control records.

Design for manufacture and assembly (DFMA) seeks to reduce unnecessary parts, awkward orientations, excessive tolerances, special tooling, and manual operations. Tolerances should be no tighter than function requires because tight tolerances increase cost and may reduce yield. Conversely, loose tolerances can produce excessive clearance, leakage, vibration, or misalignment.

A prototype is not automatically a production design. Prototype parts may be hand-fitted, made from unusually high-quality material, or inspected more carefully than production parts. Production validation should use representative processes, equipment, operators, materials, and inspection systems.

Process capability and quality control are especially important for safety-critical dimensions. Measurement systems must themselves be suitable: an inaccurate gauge can create false confidence. Nonconformances should be investigated for systemic causes rather than corrected only by sorting defective parts.

Digital tools support this transition. Computer-aided design, finite-element analysis, tolerance analysis, digital manufacturing records, and product-lifecycle management can improve traceability. Artificial intelligence may assist with alternative generation, anomaly detection, document search, or predictive maintenance, but engineering responsibility remains with qualified personnel. AI-generated geometry or recommendations require independent verification, clear data provenance, and controls against hidden assumptions, biased training data, and untraceable decisions.

10. Illustrative Case: Redesign of a Conveyor Drive Shaft

This fictional case illustrates the design reasoning process.

Background

A packaging conveyor uses a motor and reducer to drive a steel roller through a shaft and keyed hub. Operators report occasional key failures during product jams. The original shaft was designed for normal running torque only.

Problem definition

The shaft must transmit 500 N·m continuously at 60 rpm. A jam can impose 900 N·m for up to 5 s. The design must fit the existing bearing centers, permit key replacement, and avoid hazardous release of the roller. The machine operates in a dusty environment and is expected to provide 10,000 operating hours.

Analysis

The team maps the load path: reducer output, coupling, shaft, key, hub, roller, bearings, housing, and frame. It identifies that the key is not merely a torque carrier; it is also a stress concentrator and a replaceable interface. The jam condition creates a transient torque well above normal operation.

At 60 rpm and 500 N·m, ideal transmitted power is:

The team does not use power alone to size the shaft because the jam torque governs strength and the repeated start-stop operation may govern fatigue. It checks shaft torsional stress, bending from roller and belt forces, key bearing and shear stresses, shoulder stress concentration, bearing reactions, and coupling alignment. It also evaluates whether a torque-limiting device can prevent recurrence.

Approach and solution

Three concepts are compared: a larger keyed shaft, a splined connection, and a torque limiter combined with a moderately larger shaft. The torque limiter has the greatest benefit because it controls the overload at its source. The selected design uses a standardized shaft diameter compatible with the existing bearings, an improved hub interface, a specified key material and fit, and a limiter set below the damaging jam torque but above normal operating torque.

The design review identifies additional controls: a guard over the rotating coupling, an inspection interval for key wear, a jam-clearing procedure, and verification of limiter operation. The shaft is tested under normal duty and controlled overload, while the assembly is checked for alignment, temperature, vibration, and safe containment.

Outcome

The redesigned system meets continuous torque requirements and prevents repeated overload transmission. The result is not simply a stronger shaft; it is a better load-path design with overload control, maintainability, and verified interfaces.

Learning points

1. The nominal operating load may not govern design.
2. A local component failure may originate from a system-level overload.
3. Functional decomposition reveals alternatives to increasing component size.
4. Standard interfaces and maintainability should be considered together.
5. Verification must include realistic transient conditions.

11. Professional Responsibilities, Safety, and Ethics

Engineers have a responsibility to protect users, workers, the public, and the environment. This includes reporting known hazards, refusing to conceal uncertainty, preserving design records, and escalating risks that cannot be justified technically. Cost and schedule are legitimate constraints, but they do not override mandatory safety requirements or honest communication.

Safety-critical design should favor inherently safer solutions, guarding, interlocks, fail-safe behavior, energy isolation, and clear warnings. Administrative instructions should not be the only control when a physical safeguard is feasible. Regulatory compliance is a minimum obligation; ethical design may require a higher level of protection when foreseeable misuse or vulnerable users are involved.

Sustainability is also a design consideration. Material selection, energy efficiency, repairability, disassembly, recyclability, and life-cycle impacts can change the preferred concept. A lightweight component that fails early may have a greater environmental burden than a durable, repairable alternative.

Challenges and Limitations

No design model perfectly represents reality. Material data may vary, loads may be uncertain, boundary conditions may be idealized, and manufacturing variation may be underestimated. Numerical simulation can produce precise-looking results from incorrect assumptions. Standards may not cover unusual configurations. Prototypes may not reproduce long-term wear or rare accidents.

The appropriate response is not to abandon analysis, but to combine methods: conservative assumptions where justified, sensitivity studies, physical tests, independent review, inspection, monitoring, and feedback from service. Engineers should state what a calculation proves, what it does not prove, and what evidence is still required.

Chapter Summary

Machine component design converts a need into a product that performs safely and reliably within technical, manufacturing, economic, environmental, and regulatory constraints. Effective design begins with measurable, traceable requirements rather than premature component selection. Functional decomposition clarifies what the system must do, while physical architecture and interface definition show how components realize those functions. Load-path analysis ensures that forces, torque, reactions, and environmental effects are followed through the complete machine.

Standards and codes support interoperability, safety, and consistent practice, but their applicability must be verified. Factors of safety provide a way to address limited uncertainty, yet they do not replace correct failure models or complete load cases. Design is iterative: concepts are analyzed, reviewed, prototyped, verified, validated, manufactured, and improved using production and field evidence. Design reviews, traceability, DFMA, ethical judgment, and disciplined change control are essential to the transition from concept to production.

Key Terms

Design: The process of defining a solution that satisfies stated needs and constraints.

Machine component: A distinct part or subassembly that contributes to machine function.

Requirement: A condition or capability that a product must satisfy.

Constraint: A limitation on possible design solutions.

Design specification: A measurable statement of performance, geometry, environment, or acceptance.

Functional decomposition: Breaking an overall purpose into smaller functions and subfunctions.

Interface: A mechanical, energy, information, environmental, or human connection between elements.

Load path: The route by which force, torque, energy, or heat is transmitted.

Failure mode: The manner in which a component or function can fail.

Verification: Demonstration that specified design requirements have been met.

Validation: Demonstration that the product satisfies the intended real-world need.

Standard: An agreed technical document defining requirements, methods, dimensions, or practices.

Code: A technical rule set that may be legally or contractually enforceable.

Factor of safety: The ratio of a limiting capacity to a design demand.

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

Maintainability: The ease and time with which a product can be inspected, repaired, or restored.

DFMA: Design for manufacture and assembly.

Traceability: The ability to connect requirements, decisions, analyses, tests, and records.

Design review: A structured evaluation of design evidence, risks, and readiness.

Configuration control: Management of approved product definitions and changes.

Review Questions

Short Answer Questions

1. Distinguish verification from validation.
2. What is the purpose of functional decomposition?
3. Define a load path and give one machine example.
4. Why is a factor of safety not a complete reliability measure?
5. State two benefits of using standardized components.

Long Answer Questions

1. Explain how a user need is transformed into a traceable design specification. Include examples of measurable and non-measurable statements.
2. Describe an iterative machine design process from concept generation through production release.
3. Explain how interfaces influence machine reliability, safety, and maintainability.
4. Discuss the role of standards, codes, and regulations in component design, including their limitations.
5. Explain why design reviews should include manufacturing, service, safety, and quality personnel as well as design engineers.

Analytical / Application Questions

1. A shaft transmits 400 N·m at 1,200 rpm. Calculate ideal power using P = 2π(N/60)T. State the assumptions and explain why the result alone is insufficient for shaft design.
2. A bracket carries a nominal load but fails after repeated operation. Develop a load-case and failure-mode investigation that considers fatigue, stress concentration, attachment, environment, and manufacturing variation.
3. Propose and justify three alternative concepts for protecting a conveyor drive from jam torque. Compare them using safety, cost, maintainability, manufacturability, and verification evidence.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes verification from validation?

A. Verification checks whether specified requirements were met; validation checks whether the intended need was satisfied.

B. Verification is performed only after production; validation is performed only during concept design.

C. Verification concerns cost; validation concerns material selection.

D. Verification replaces testing; validation replaces analysis.

Correct Answer: A. Verification checks whether specified requirements were met; validation checks whether the intended need was satisfied.

Question 2

What is the primary purpose of functional decomposition?

A. To eliminate the need for standards

B. To break an overall purpose into manageable functions and subfunctions

C. To select a material before defining requirements

D. To increase the number of components in an assembly

Correct Answer: B. To break an overall purpose into manageable functions and subfunctions

Question 3

A shaft transmits 500 N·m at 60 rpm. What is its ideal transmitted power?

A. 0.52 kW

B. 3.14 kW

C. 18.85 kW

D. 31.4 kW

Correct Answer: B. 3.14 kW

Question 4

Which item is part of a complete load-path analysis?

A. Only the component with the highest stress

B. Only the nominal operating load

C. Applied loads, reactions, interfaces, and supporting structure

D. Only the purchased components

Correct Answer: C. Applied loads, reactions, interfaces, and supporting structure

Question 5

For a simple tensile member, the nominal normal stress is calculated as:

A. σ = A/F

B. σ = F/A

C. σ = FL/4

D. σ = 16T/(πd³)

Correct Answer: B. σ = F/A

Question 6

What is a major limitation of using a very large factor of safety?

A. It guarantees validation under all conditions

B. It may add unnecessary mass and cost without correcting an omitted failure mode

C. It eliminates the need for standards

D. It makes fatigue impossible

Correct Answer: B. It may add unnecessary mass and cost without correcting an omitted failure mode

Question 7

Which document most directly connects a requirement to its analysis and test evidence?

A. Requirements traceability matrix

B. Supplier invoice

C. Lubrication chart only

D. Assembly photograph

Correct Answer: A. Requirements traceability matrix

Question 8

Why should tolerances generally not be tighter than function requires?

A. Tight tolerances always reduce accuracy

B. Tight tolerances usually increase manufacturing and inspection cost

C. Loose tolerances prevent all failures

D. Tolerances apply only to cosmetic features

Correct Answer: B. Tight tolerances usually increase manufacturing and inspection cost

Question 9

Which concept most directly addresses an overload at its source in the conveyor case?

A. Increasing paint thickness

B. Adding a torque limiter

C. Reducing the bearing temperature rating

D. Removing the protective guard

Correct Answer: B. Adding a torque limiter

Question 10

A standard is best described as:

A. Always a criminal law

B. An agreed technical document defining requirements, methods, dimensions, or practices

C. An informal personal preference

D. A guarantee that every design using it is safe

Correct Answer: B. An agreed technical document defining requirements, methods, dimensions, or practices