To engineer is human: the role of failure in successful design

『To engineer is human』の要約

1. 1

   When Failure Becomes Our Greatest Teacher

   再生を開始

   Every day, we cross bridges, enter buildings, and board airplanes without a second thought. Yet behind each structure lies a fascinating paradox: the engineers who create these marvels spend most of their time thinking about how they might fail. Henry Petroski's "To Engineer Is Human" has become a cult classic among both engineering professionals and the intellectually curious public since its 1985 publication. Bill Gates included it in his "5 books to read this summer" list in 2015, praising how it transformed his thinking about innovation. The book's enduring appeal comes from Petroski's remarkable ability to reveal the human dimension behind technical achievements-showing that engineering isn't merely about mathematical precision but about understanding human limitations, learning from mistakes, and constantly improving through failure. As our infrastructure ages and climate challenges intensify, Petroski's insights about how we learn more from collapse than success have never been more relevant to our collective future.

2. 2

   The Paradox of Engineering and Human Nature

   再生を開始

   When disasters strike-like the 1981 Kansas City Hyatt Regency walkway collapse that killed 114 people-we instinctively question how engineers could have allowed such catastrophes. Yet engineering, at its core, is profoundly human. Unlike nature's designs, refined over millions of years, human-made structures constantly evolve with our changing tastes, resources, and ambitions. We demand fashionable buildings and pursue ever-larger creations, making engineering both exciting and perilous. Despite technological advancement, recent decades have witnessed some of history's costliest structural failures. The concern isn't new-the Code of Hammurabi from 4,000 years ago included harsh penalties for builders whose houses collapsed. Modern consequences differ: the Hyatt disaster triggered billions in lawsuits rather than executions, with investigations continuing years afterward. What's striking is that structural failures make headlines precisely because they're extraordinarily rare. The probability of a building collapse ranges from one in a million to one in a hundred trillion. While automobile accidents claim tens of thousands of lives annually, we notice engineering failures because they're dramatic anomalies. Our relationship with technology evolves from fear to acceptance-just as air travel transformed from terrifying to routine. The reality is that all structures could be built ten times stronger, but at tremendous cost. We constantly balance safety with economics, making calculated compromises. Less-developed countries often experience more frequent structural failures because they lack resources for overdesign. Meanwhile, America's aging infrastructure requires trillions in repairs, as demonstrated when a century-old water main broke in Manhattan in 1983, causing electrical outages and millions in business losses. Engineering isn't just about preventing failure-it's about managing acceptable risks within human constraints.

3. 3

   The Playground of Structural Understanding

   再生を開始

   We're all born engineers. From infancy, we exist in a world balanced between trust and risk, learning the fundamentals of structural engineering through experience. Our bodies function as complex structures-levers, beams, columns-as we learn to roll over, crawl, and eventually walk. Each fall teaches us what not to do next time. We don't so much learn to walk as learn not to fall. Even nursery rhymes prepare us for structural realities. "London Bridge is falling down" teaches both failure and reconstruction. "Ring around the rosie" and "Step on a crack" introduce structural vulnerability. Through play with blocks, Erector Sets, and Legos, children develop an intuitive understanding of structural principles that serves them throughout life. The anthropomorphic language of engineering reflects this bodily experience-both humans and structures experience stress, strain, fatigue, and collapse. We understand intuitively that chairs and beds, like people, have strength limitations. Fairy tales like "The Three Little Pigs" demonstrate how different design choices yield different outcomes under extreme conditions. I demonstrate fatigue failure to my engineering students with simple paper clips. By repeatedly bending them until they break, I show how cyclic stress leads to failure. The number of bends before breaking varies because not all clips are equally strong and students bend them differently. This classroom experiment produces a classic bell-shaped distribution curve, teaching students that fatigue failure is predictable but not precisely so. Many everyday annoyances stem from this fatigue phenomenon. Light bulbs fail not from continuous use but from repeated on-off cycles that stress the filament. Children's toys are particularly susceptible since they're not overdesigned-making them too rugged would make them too heavy and expensive. My son's Speak & Spell electronic toy provided a perfect case study when its keys began breaking. The correlation between broken keys and the most frequently used letters in English was striking. The ENTER key broke first (used for every word), followed by E, T, O, S, and R-five of the seven most common letters in English. This wasn't coincidental but a predictable pattern of fatigue failure where the most-used components break first. Despite this childhood preparation for the reality of structural failure, as adults we paradoxically expect our engineering achievements to be immortal. We want our creations to transcend human limitations, forgetting the lessons of our youth that taught us everything eventually breaks.

4. 第4章

   The Engineering Hypothesis

   The British Institution of Structural Engineers defines their field as "the science and art of designing and making, with economy and elegance, buildings, bridges, frameworks, and other similar structures so that they can safely resist the forces to which they may be subjected." This definition embraces engineering as both science and art-requiring imaginative conception like an artist and rigorous analysis like a scientist. It emphasizes economy (not wasting resources), elegance (professional excellence), and above all, safety. Science constructs theories about natural behavior, combining observation, experience, and inspiration. While theory creation follows no universal method, the scientific method tests hypotheses by comparing conclusions with reality. Verification never proves a hypothesis true, as one untested case might disprove it, but a single contradiction renders it false. Even seemingly established "facts" remain hypotheses-confirmed repeatedly but never proven. Engineering design resembles scientific theorizing, but engineers hypothesize about their own creations rather than the natural world. Each new bridge or building represents a hypothesis that it will stand without failing under specified conditions. A bridge standing for years confirms but never proves this hypothesis, while sudden collapse disproves it completely. Design involves successive hypotheses that arrangements of parts will perform without failure, each tested through analysis. Designers strengthen weak links until no imaginable failure remains, though overlooked calculations or questions can lead to false verification. Engineering failures-like the Hyatt Regency walkways, Tacoma Narrows Bridge, or Teton Dam-represent disproved hypotheses. Even the Brooklyn Bridge's century-long success only confirms its past reliability, not its future performance. We accept such uncertainty as we do health risks-acknowledging remote dangers while functioning in an imperfect world. The beam-spanning space while resisting bending forces-exemplifies fundamental structural forms. Galileo pioneered engineering science by examining cantilever beams, correctly observing that beam strength is proportional to the square of depth. However, he made mistaken assumptions about force distribution, arriving at qualitatively correct but quantitatively imprecise answers-highlighting why engineering demands rigorous checking. As structures grow more complex than simple beams, intuition becomes insufficient, requiring formal analysis as the profession's lingua franca. The engineering-scientific method becomes the arbiter of competing conclusions, ensuring that structures stand not through hope but through careful verification of each hypothesis about their behavior.

   第5章

   Success Through Anticipating Failure

   Engineering success begins with anticipating failure-not just catastrophic collapse but also structural unsoundness like corrosion, service breakdowns, and fatigue cracking. A designer who overlooks even one failure mode risks everything, regardless of how ingenious the conception appears. The Egyptian pyramids exemplify early engineering through trial and error. Their stable shape evolved from low rectangular mastabas to stepped structures and finally to true pyramids. The Bent Pyramid at Dahshur dramatically changes slope midway up-likely evidence of structural failure forcing designers to adjust their ambitions. Subsequent builders constructed higher but not steeper pyramids, learning from this cautionary example. Gothic cathedrals evolved through similar experimentation, with builders learning from neighbors' successes and failures. The 1284 collapse at Beauvais Cathedral marked a turning point, prompting more conservative approaches. Flying buttresses and decorative pinnacles were functional responses to cracking masonry, not merely aesthetic choices. Nineteenth-century railroads transformed landscapes and engineering practices. As engineering evolved from trial-and-error to scientific calculation, engineers became seen as those applying "scientific process" to structural problems. They developed methods to calculate and prevent material failures, though they couldn't eliminate mental oversights. Engineers, contrary to their conservative reputation, constantly push boundaries-always believing they're avoiding errors while actually conducting new trials with each structure. The public remains threatened by both their failures and successes-such is the nature of all human endeavors. This tension between innovation and caution defines engineering's essential character.

   第6章

   From Concept to Construction: The Design Journey

   Designing a bridge resembles planning a vacation-the destination is clear but the means are limited only by imagination. Just as families choose between driving or flying based on economics, convenience, aesthetics, emotions, and habit, engineers face similar choices in design. Both processes involve countless decisions, some trivial and others critical to success. The best approach often involves anticipating what could go wrong-families avoid restaurants that disappointed them previously, while engineers anticipate potential failures. Engineering design draws on accumulated experience, yet sometimes demands innovation beyond precedent. Modern bridges evolved from primitive logs across brooks to sophisticated steel and concrete structures designed for specific functional, aesthetic, and economic demands. When spanning wider bays or deeper ravines than ever before, designers cannot simply copy past successes. The first iron bridge at Coalbrookdale in 1779 represented a bold experiment with new materials, mimicking familiar stone-and-timber construction details. Once iron proved viable, engineers began exploiting its tensile properties in ways stone could never match, necessitating a period of trial and error. The railroads' expansion demanded bridges, initially built from familiar timber that required maintenance and was vulnerable to fire. Iron gradually replaced wood despite early unpredictability and numerous collapses. Famous failures like the Tay Bridge (1879) and Quebec Bridge (1907) punctuate bridge history, but countless unnamed railroad bridge collapses during the nineteenth century raised public concern about the new technology. Innovation in engineering involves unpredictable outcomes. Each new design opportunity presents countless choices-whether to copy successful features from existing structures or improve upon their apparent weaknesses. A bridge that has stood for decades but developed minor cracks might inspire an improved design, while a trouble-free structure might suggest opportunities for lighter, less expensive construction. These design choices parallel life choices. Though engineers may develop multiple designs on paper, ultimately only one can be built, just as only one route can be taken on a journey. As Robert Frost wrote about diverging roads, engineers must peer into an uncertain future where paths continually fork. Conservative engineers may choose familiar, proven designs despite higher costs or aesthetic compromises. Bold innovators like Swiss engineer Robert Maillart revolutionize their field by taking "the road less traveled," creating structures of both economy and beauty that make "all the difference."

   第7章

   Revisions and Refinements: The Creative Process

   The writer staring at a blank page surrounded by crumpled false starts mirrors the engineering design process. Like authors arranging words to take readers from here to there, engineers discard unsatisfying attempts based on judgments of what works. Writers preserve drafts and false starts, recognizing they'll never achieve perfection but must eventually choose the least imperfect attempt. These documents reveal creation as a succession of choices-an opening sentence might evolve through dozens of rejected alternatives before reaching its final form. The creative process is more easily understood through rejection than acceptance, through failure rather than success. Masters like James Joyce were never completely satisfied, making changes even as works were being typeset. Engineers similarly abandon rather than complete designs, reaching a point of diminishing returns where they've caught major flaws but recognize perfection is unattainable. The engineer, like the writer, eventually reaches a version believed to be as free of flaws as possible before submitting it to others for assessment. Both writing and engineering involve successive elimination of faults and errors, with great practitioners in both fields learning more from mistakes than successes. This iterative process of refinement distinguishes truly excellent design from merely adequate solutions. When engineers face competing demands like fuel efficiency, light weight, accessibility, and comfort, radical design changes can lead to unexpected failures. The Grumman Flxible buses purchased by New York City demonstrate this principle perfectly. Despite their innovative lightweight frames designed to meet federal specifications, these buses developed severe fatigue cracks after just months of service, forcing the entire 637-bus fleet off the road. This engineering failure, while causing no catastrophic accidents, reveals the dangers of departing too dramatically from proven designs. The composition of materials adds another layer of complexity-austenitic stainless steel, for instance, contains precise proportions of iron, chromium, nickel, and trace elements that dramatically affect performance. Even small variations in manufacturing can produce dramatically different results in seemingly identical products.

   第8章

   Learning from Catastrophe: Case Studies in Failure

   On July 17, 1981, the Kansas City Hyatt Regency Hotel became synonymous with America's greatest structural tragedy when two crowded walkways collapsed onto the lobby below, killing 114 people and injuring nearly 200 others. The catastrophe occurred during a Friday evening event where dancers and spectators filled the architecturally graceful walkways suspended across the hotel's columnless atrium. The original design called for single long rods extending from the ceiling through both walkways. Though underdesigned at only 60% of code requirements, this configuration might have survived. However, someone proposed "a better idea"-using two shorter rods instead of each long one. This seemingly practical change for easier installation proved catastrophic, effectively doubling the load on the upper walkway's connection points. Multiple warning signs preceded the collapse. The lobby roof had collapsed during construction, prompting checks of structural details that somehow missed the critical rod connections. Construction workers reported the walkways feeling unsteady under heavy loads, but traffic was merely rerouted rather than the design being scrutinized. After twenty months of investigation, prosecutors found no evidence of criminal wrongdoing, though Missouri's attorney general later filed complaints against the engineers for "gross negligence" in failing to calculate load capacities and verify the modified design. The tragedy highlighted the critical importance of alternate load paths in structural design-redundancies that allow structures to redistribute weight when one support fails. Similarly, the Mianus Bridge collapse on Interstate 95 in Connecticut revealed how critical detail failures without redundant load paths can lead to disaster. The bridge was a "pin-connected hung span" that crossed the river at an angle. When one connection failed due to corrosion from salt-contaminated rainwater, the entire span collapsed. These failures-the Hyatt skywalks, Mianus Bridge, and others-all involved critical detail failures without redundant load paths. Rather than dismantling all such structures, engineers must study these failures to identify and strengthen weak links in existing designs and avoid them in future ones. The positive lesson from failure is what not to do, making even tragic structural collapses valuable learning opportunities.

   第9章

   The Crystal Palace: A Triumph of Innovation

   Innovation in engineering inherently involves risk, but doesn't necessarily lead to failure. The Crystal Palace, built for London's Great Exhibition of 1851, exemplifies how bold innovation can triumph despite opposition. Joseph Paxton's design faced numerous challenges-from Parliament debates about foreign competition to passionate defense of a small clump of elm trees. Despite entering the competition late, Paxton's "ferro-vitreous" concept ultimately won unanimous support from the building committee due to its simplicity, speed of assembly, absence of internal walls, and material reusability. The design's mathematical regularity simplified construction-with 24-foot units becoming the basic scale, determined by practical glass manufacturing constraints. Workers installed glass using trolleys on "Paxton gutters" that doubled as drainage systems. The building's elegant proportions came from this modular approach, with deep girders and trusses providing both aesthetic grace and structural rigidity. Despite predictions of catastrophic failure from mathematicians and engineers, the Crystal Palace incorporated extensive safety measures. All cast-iron girders were hydraulically tested, diagonal rods provided lateral stiffness, and the galleries underwent extreme load testing-supporting 300 workmen jumping simultaneously with minimal deflection. Queen Victoria opened the Great Exhibition on May 1, 1851, and over six million people visited during its 141-day run, with up to 90,000 people in the building simultaneously without incident. The Crystal Palace represented a revolutionary breakthrough in construction despite criticism from traditional architects and engineers who questioned Paxton's lack of formal training. The Palace's influence continues in contemporary architecture, from the curtain walls of modern skyscrapers to exposed structural elements in public buildings. Its success demonstrates that innovation, when approached with proper care and testing, need not lead to failure but can instead create revolutionary breakthroughs that influence generations to come.

   第10章

   The Balance of Safety and Innovation

   While engineers learn what not to do from failures, successes only teach them how to repeat the same design-which can be dangerous when conditions change. Each engineering project carries potential for failure, which is why the factor of safety is crucial. Often called the "factor of ignorance," it provides margin for error by dividing the load required to cause failure by the maximum expected load. Engineers must make structures tough rather than fragile, while balancing strength against aesthetics, economics, and necessity. When structures fail despite safety factors, engineers typically increase the factor for similar future designs. Conversely, when structures consistently succeed, designers often reduce the "factor of ignorance" as confidence grows. This cyclic pattern of raising safety factors after accidents and lowering them during safe periods can lead to recurring failures. Bridge design particularly illustrates engineering's delicate balance between success and failure. The paradox of engineering design is that successful concepts often devolve into failures, while colossal failures drive innovation. When we understand design's primary objective as preventing failure, this contradiction resolves itself. The interplay between success and failure is nowhere more dramatic than in suspension bridge history. John Roebling's Brooklyn Bridge succeeded because he understood the importance of properly stiffened decks. However, later engineers gradually made roadways longer and less deep, creating increasingly flexible structures that culminated in the spectacular failure of the Tacoma Narrows Bridge. Historical precedents should have warned engineers about such dangers. The Wheeling Suspension Bridge was destroyed by wind after five years. A reporter's dramatic account described how the bridge "lunged like a ship in a storm" before twisting and crashing into the river. These lessons went unheeded by modern designers. While failures teach caution, successes inspire innovation. We could eliminate structural failures by halting innovation altogether, but this would mean no bridges could be built where none existed before. Just as great art inspires new creation rather than mere copying, great bridges urge engineers to push boundaries responsibly.

   第11章

   From Slide Rules to Computers: The Changing Face of Engineering

   Twenty-five years ago, the slide rule was engineering's undisputed symbol. Students carried these "slip sticks" in scabbard-like cases on their belts, while older engineers wore miniature working models as tie clips. These analog calculators taught crucial engineering principles: answers are approximations that should only be reported to appropriate precision, and magnitudes come from understanding problems, not automatically from calculating devices. By the mid-1970s, calculators had displaced slide rules, and eventually computers transformed engineering practice entirely. Computers enable engineers to tackle previously impossible problems, particularly in structural analysis. Complex structures that would have required prohibitive calculation time can now be optimized through software. However, if errors exist in translating a design concept into a numerical model, the computer's output might bear little relation to reality. Since engineers may lack intuitive feel for such complex structures, they might not recognize suspicious results. Computer programs have been found with fundamental errors, like incorrect values for pi or sign errors that understated earthquake stresses in nuclear plant designs. Computer optimization creates structures where every component has minimum weight and strength for maximum economy-a degree of efficiency impossible with manual calculations. However, this leaves little margin for error in calculations, manufacturing, or construction. Tests comparing computer predictions with actual structural behavior have shown disturbing inaccuracies, with predictions within only 60% of measured values 95% of the time. While computers excel at analyzing familiar designs where engineers know what questions to ask, they become dangerous tools when designing innovative structures without precedent. As we create increasingly complex designs because we believe computers can do what humans cannot, failures become more likely-for the further we stray from experience, the less likely we are to anticipate all the right questions. The most catastrophic failures typically occur amid overconfidence, not paranoia. The 1986 Challenger disaster exemplifies this-despite engineers' concerns about O-rings functioning properly in unusually cold weather, their warnings were dismissed by managers who pointed to the perfect launch record. Only after disaster did NASA give engineers' doubts proper weight. Engineering is fundamentally human-we must embrace both innovation and caution, learning from history's lessons. Successful innovations like the Iron Bridge, Eads Bridge, and Brooklyn Bridge demonstrate that cautious exploration with new materials can succeed brilliantly. Yet history also warns that materials in new environments can lead to cracked nuclear plants, fast-track construction can fail spectacularly, and computers don't automatically produce better designs. The essential lesson is that innovation inherently involves risk. Engineers who prepare thoroughly for innovation deserve no more blame for failure than a well-trained pole vaulter who misses a record attempt. It's the unprepared engineer, or one who repeats known mistakes, who acts irresponsibly. By openly sharing discovered errors, engineers increase their collective ability to catch potentially dangerous mistakes before they cause harm. To engineer is human-to learn from failure, divine.

   続きを読む