Machine Design

Machine design is the branch of mechanical engineering that decides what a machine's parts should be: their shape, material, size, fits, and arrangement, so that the machine transmits motion, carries loads, and survives years of use without failing or falling apart.

12 frameworks, from 100 to today · Mechanical Engineering · Engineering

still a working view ended · each leads to its entry below

Overview

Machine design is the branch of mechanical engineering that decides what a machine's parts should be: their shape, material, size, fits, and arrangement, so that the machine transmits motion, carries loads, and survives years of use without failing or falling apart. It sits between analysis and manufacture. The analyst asks whether a beam will break; the designer decides how thick the beam should be, what it should be made of, how it attaches to its neighbour, and how the whole assembly will be built and repaired. What is at stake is concrete: a gearbox that runs quietly for twenty years, a crane hook that does not crack, a pump whose shaft does not wobble itself to destruction, a product that can be assembled without error by a worker on a line.

Craft Treatises and the Repertoire of Mechanisms

Before there was a science of machine design there was a literature of machines. From antiquity through the Renaissance, engineers compiled treatises describing devices they had seen or built: water-lifting machines, mills, pumps, cranes, and military engines. The tradition is best represented by the Pneumatica attributed to Hero of Alexandria (1st century CE), which describes siphons, force pumps, and a rotating steam device, and by the De ingeneis and De machinis of the Sienese engineer Mariano di Jacopo, called Taccola (early 15th century), and the notebooks of Francesco di Giorgio Martini, which circulated widely in manuscript.

The decisive turn came with the printed book. Agostino Ramelli's Le diverse et artificiose machine (1588) presented 195 engravings of machines, including his famous book wheel, and Vittorio Zonca's Novo teatro di machine et edificii (1607) did similar work for mill and pump machinery. These books were pattern books: a designer facing a problem looked up a device that solved something similar and adapted it. Their lasting contribution was a repertoire—the crank, the cam, the ratchet, the screw, the linkage—that later designers could combine. What they could not do was predict whether a chosen device would work or how strong it needed to be; that judgement remained with the craftsman.

Kinematic Geometry: Describing Motion Before Building It

The first genuinely theoretical strand of machine design was kinematics, the geometry of motion considered apart from the forces that cause it. Gaspard Monge's descriptive geometry, taught at the École Polytechnique from the 1790s, gave engineers a systematic way to represent three-dimensional shapes on paper. His student Jean-Victor Poncelet, in his Cours de mécanique appliquée aux machines (1826 and later), set out the geometry of toothed gearing and of mechanisms in a form engineers could use.

The classic statement of the field is Robert Willis's Principles of Mechanism (1841), which classified mechanisms by the motion they produced and introduced the analysis of toothed wheels by their pitch surfaces. Franz Reuleaux, at the Berlin trade school that became the Technical University, went further. His Theoretische Kinematik (1875; English translation The Kinematics of Machinery, 1876) treated a machine as a chain of rigid bodies connected by joints, and introduced the idea of the kinematic pair—two bodies in contact that constrain each other's relative motion. Reuleaux's collection of several hundred mechanism models, built to teach this analysis, survives at Cornell University and elsewhere. The payoff was that a designer could now reason about a linkage's motion on paper, count its degrees of freedom, and predict whether a proposed arrangement would jam or move as intended, before cutting any metal.

Sizing for Strength: From Rule of Thumb to Calculation

Parallel to kinematics, a science of strength emerged. Galileo's Two New Sciences (1638) analysed a cantilever beam and got the stress distribution wrong, but posed the right question. The correct theory of bending came from Claude-Louis Navier's Résumé des leçons (1826), which gave the now-standard formula relating bending moment, second moment of area, and stress, and from Siméon Denis Poisson and others on elasticity. Navier's work, taught at the École des Ponts et Chaussées, turned shaft and beam sizing from a craft rule into a calculation.

The practical consequence was the rise of the handbook. William Fairbairn's experiments on wrought-iron beams in the 1840s, and his Treatise on Mills and Millwork (1861–63), gave designers empirical data on real materials. By the late 19th century, engineers worked from tables of allowable stress for cast iron, wrought iron, and steel, and sized shafts, keys, and bolts by simple formulas with generous safety factors. The safety factor was the honest admission that the calculation was approximate: if the load was uncertain and the material variable, design for several times the expected stress.

The Drawing Office and the Birth of the Design Process

As calculations grew, so did the organization that carried them out. The 19th-century drawing office, staffed by draughtsmen working at tilted boards, became the place where a machine was actually designed. Orthographic projection—front, plan, and side views to scale—allowed a complex assembly to be specified unambiguously, and the drawing became the contract between designer and workshop. Standards followed: screw threads were standardized in Britain from the 1840s (Joseph Whitworth's thread of 1841), and national standards bodies in the early 20th century codified dimensions, material specifications, and drawing conventions.

This is also where limits and fits entered design. A shaft must be slightly smaller than its hole to assemble, but how much smaller? Before standardization, each shop had its own practice, and parts from different shops would not interchange. The tolerancing systems formalized in the early 20th century, and later the ISO system of limits and fits, defined classes of clearance, transition, and interference fits so that a designer could specify "this bearing must press on this shaft" in a way any competent shop could reproduce. Interchangeability, the basis of mass production, depended on this.

Machine Elements: Designing the Parts That Recur

By the early 20th century, designers recognized that most machines are built from a small set of recurring components: shafts, bearings, gears, belts, chains, couplings, clutches, brakes, springs, fasteners, and seals. The discipline of machine-element design grew up around these. Carl Barth's work on spindle design and the tapered roller bearing patented by Henry Timken in 1898 are examples of the component-level thinking that came to dominate. Textbooks such as those by Robert Thurston and later by Joseph Shigley (whose Mechanical Engineering Design, first published in 1963, became a standard teaching text) organized the field around the analysis and selection of these elements.

The logic of machine-element design is that each element has a characteristic failure mode and a characteristic calculation. A shaft fails by fatigue at a stress concentration; a ball bearing fails by rolling-contact fatigue and is rated by its basic dynamic load rating; a bolted joint fails by separation or by bolt fatigue. Designing a machine became, in large part, the art of choosing and sizing standard elements correctly and arranging them so that each is loaded in the way its rating assumes.

Fatigue, Fracture, and the Limits of Static Calculation

The most important correction to strength-of-materials sizing came from the study of fatigue: failure under repeated stress at levels well below the static strength. The phenomenon was investigated systematically by August Wöhler in the 1850s and 1860s, working for the Prussian railways on axle failures, who established the S–N curve (stress versus number of cycles to failure) and the concept of an endurance limit. The lesson was that a component designed to static allowable stress could still break after a million cycles.

The 20th century added stress-concentration analysis (the recognition that a notch or a sharp change of section multiplies local stress), cumulative-damage rules such as the Palmgren–Miner linear rule, and, from the 1950s, fracture mechanics, which treats a crack as a stress raiser with its own governing parameter, the stress-intensity factor. The practical effect on design was profound: fillets, generous radii, surface finish, and shot-peening became design decisions with quantitative justification, and the safety factor was increasingly replaced by explicit fatigue and fracture assessments.

Dynamics and Balancing: When Motion Itself Is the Problem

Rotating and reciprocating machinery brought a class of problems that statics could not touch. An unbalanced rotor shakes its bearings; a reciprocating engine shakes its frame; a shaft spinning near a critical speed whips violently. Rankine's analysis of whirling shafts in the 1860s, and later work by Henry Jeffcott and others in the early 20th century, explained the critical speed as a resonance between rotation and the shaft's own bending natural frequency. Balancing—adding or removing mass so that the centre of gravity lies on the axis—became a standard workshop and design operation, with balancing grades eventually standardized.

The same period saw the systematic study of flywheels for smoothing torque fluctuations, of governors for speed control, and of vibration isolation. The design office acquired a dynamics calculation alongside its strength calculation.

Systematic Design Methodology

After the Second World War, a movement arose to make design itself a teachable, systematic process rather than a matter of individual flair. In West Germany, Gerhard Pahl and Wolfgang Beitz published Konstruktionslehre (1977; English Engineering Design: A Systematic Approach, 1984), which divided design into clarifying the task, conceptual design (abstracting the problem to its essential functions and searching for working principles), embodiment design, and detail design. In Britain, the Design Research Society and the conferences of the 1960s and 1970s pursued similar goals. The claim was that a structured method would produce better and more original solutions than jumping straight to a familiar shape.

The method's most durable legacy is functional decomposition: describing what a machine must do as a set of functions (transmit torque, convert rotation to translation, seal against pressure) and then seeking a physical principle for each. It also gave design education a vocabulary—requirements list, morphological chart, evaluation matrix—that persists in engineering curricula.

Reliability, Manufacture, and the Statistical Turn

As products grew more complex and customers less tolerant of failure, design began to treat failure statistically. Reliability engineering, developed largely in the US aerospace and electronics industries from the 1950s, asked for the probability that a component would perform its function for a stated time. Weibull's distribution, published in 1951, became a standard tool for describing the scatter in fatigue life and strength. Design for reliability meant identifying failure modes systematically (fault-tree analysis, failure modes and effects analysis) and designing them out or providing redundancy.

A parallel movement addressed manufacture. Design for manufacture and assembly (DFMA), associated with Geoffrey Boothroyd and Peter Dewhurst from the 1960s onward, quantified the cost of assembly—how many parts, how many fasteners, how many directions of approach—and pushed designers toward fewer, simpler parts. The insight was that most of a product's cost is fixed at the design stage, so the designer, not the factory, controls manufacturability.

Computer-Aided Design and Numerical Analysis

The largest change of the late 20th century was computational. Interactive computer graphics for design emerged in the 1960s, notably Ivan Sutherland's Sketchpad (1963) and, in industry, the DAC-1 system developed at General Motors in the early 1960s. By the 1980s, three-dimensional solid modelling had replaced the drawing board for much design work: a part was now a mathematical solid that could be sectioned, measured, and checked for interference automatically.

Alongside it came the finite element method (FEM), which divides a complex shape into small elements and solves the governing equations numerically1. The method's mathematical foundations were laid in the 1940s and 1950s, and its engineering application to structural analysis spread through the 1960s and 1970s as computers became affordable. For machine design the consequence was that stress in a real, irregular part—a connecting rod with its fillets, a housing with its ribs—could be computed rather than estimated. Later, kinematic and dynamic simulation of whole mechanisms, and optimization routines that search for the lightest or stiffest shape, became routine. The designer's judgement did not disappear, but the calculation that once took a week of hand work could now be iterated in an afternoon.

The Present

Machine design today is a synthesis of all these strands. A designer works in a solid-modelling environment, sizes standard elements from catalogues and standards, checks fatigue and sometimes fracture, simulates motion and vibration, and evaluates manufacturability and assembly cost, often within a concurrent-engineering team where manufacturing and service engineers contribute before the design is frozen. Reliability methods and statistical tolerancing are standard where failure is costly. The open questions are familiar in new form: how to design for additive manufacturing, where the old rules of casting and machining no longer apply; how to validate simulation against test when the simulation is cheaper than the experiment; how to design machines whose sensors and controllers are as important as their structure, blurring the line with mechatronics; and how to make design decisions defensible when the loads, the material properties, and the future use are all uncertain. The craft treatise gave designers a repertoire, kinematics gave them a language for motion, strength of materials gave them calculation, the drawing office gave them a process, fatigue and reliability gave them humility about scatter, and the computer gave them speed. The task itself—deciding what to build so that it works and lasts—has not changed.

References

  1. Finite element method, Wikipedia

The frameworks of Machine Design

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 100-1650

    Craft Treatise and Mechanism Repertoire

    Which known device solves this machine problem?

    Fly to its tower in Nousfarer

  2. 1790-Present

    Kinematic Geometry

    What path does each part of a mechanism trace?

    Fly to its tower in Nousfarer

  3. 1826-Present

    Strength of materials

    How thick must a part be to carry its load?

    Fly to its tower in Nousfarer

  4. 1840-1990
  5. 1860-Present

    Fatigue

    How long will a part last under repeated stress?

    Fly to its tower in Nousfarer

  6. 1860-Present

    Rotordynamics

    Why does a spinning shaft shake itself apart?

    Fly to its tower in Nousfarer

  7. 1900-Present

    Limits and Fits

    How loose or tight should a shaft fit its hole?

    Fly to its tower in Nousfarer

  8. 1900-Present

    Machine element

    How should a standard part be chosen and sized?

    Fly to its tower in Nousfarer

  9. 1950-Present

    Reliability engineering

    How often will a product fail in service?

    Fly to its tower in Nousfarer

  10. 1960-Present

    DFMA

    How can a product be made and assembled cheaply?

    Fly to its tower in Nousfarer

  11. 1960-Present

    Systematic Design Methodology

    What steps turn a need into a working design?

    Fly to its tower in Nousfarer

  12. 1963-Present

    Computer-Aided Design and Numerical Analysis

    How can a computer model predict a design's behaviour?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Craft Treatise and Mechanism Repertoi…Kinematic GeometryStrength of materialsEngineering Drawing and Tolerancing P…FatigueRotordynamicsLimits and FitsMachine elementReliability engineeringDFMASystematic Design MethodologyComputer-Aided Design and Numerical A…

Beyond the field

Where Machine Design's ideas came from in other fields, and what they led to there: the strongest 12 of 86 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Machine Design is a planet in Nousfarer

Its city has a tower for each of these 12 frameworks. Land, walk in, and climb one concept at a time.

Land on Machine Design

Related fields

Fields its ideas are linked to

More in Mechanical Engineering

See where Machine Design sits on the Map of Human Knowledge