How Invisible Maintenance Prevents Visible Disasters

TL;DR. Today, in all likelihood, no water main broke on your street, no bridge buckled under your commute, no elevator stalled between floors, and the power stayed on. Nothing happened, and that nothing is the entire subject of this chapter. Reliability like that isn't a natural resting state that systems settle into on their own. It's bought, continuously, through three different kinds of maintenance: reactive (fix it after it breaks), preventive (replace or service it on a schedule, before it's statistically likely to fail), and predictive (use sensor data to catch a failure developing before it happens). Preventive work is, by widely cited estimates including the U.S. Department of Energy's own facility guidance, roughly three to five times cheaper than letting the same part fail and fixing it as an emergency, and yet it is routinely the first line item cut when a budget tightens, because nobody notices the disaster a maintenance crew quietly prevented. Almost every regulation named in this book (bridge inspections, grid reliability standards, elevator safety tests, facade laws) exists because of one specific disaster in the past, not because engineers anticipated the danger in the abstract. The 1967 collapse of the Silver Bridge in West Virginia, which killed 46 people, is the direct reason the United States now inspects its more than 623,000 bridges on a legally mandated schedule. And the American Society of Civil Engineers' 2025 report card still puts the country's infrastructure at a C grade overall, with a $3.7 trillion funding gap over the next decade, a reminder that a schedule existing on paper and a schedule being funded are two different things.

Key takeaways

  • Maintenance comes in three basic forms: reactive (run to failure, then repair), preventive (scheduled service based on a component's known failure pattern), and predictive (condition monitoring and sensor data used to anticipate a specific failure before it happens).
  • Preventive maintenance is consistently estimated at three to five times cheaper than reactive repair once emergency labor, rush parts, and cascading damage are counted, yet it competes for funding against visible projects that produce a ribbon-cutting, and it usually loses.
  • Asset management and reliability engineering are real, credentialed professions, not informal janitorial work; the Society for Maintenance and Reliability Professionals (SMRP) certifies practitioners through an ANSI-accredited exam covering five distinct bodies of knowledge.
  • Most of the maintenance regimes described across this book exist because of a specific past failure, not abstract foresight: the Silver Bridge collapse (46 dead, 1967) created the National Bridge Inspection Standards that now require most U.S. bridges to be checked at least every two years.
  • Federal law requires state transportation departments to keep formal, risk-based asset management plans for bridges and pavement, but having a plan and fully funding it are legally and financially separate questions, as the 2022 Fern Hollow Bridge collapse in Pittsburgh showed.
  • The 2025 ASCE Infrastructure Report Card gives the United States an overall C grade (its best ever) and estimates a $3.7 trillion gap between planned spending and what full repair would actually cost through 2033.

The day nothing happened

Nothing happened today. The tap ran clear when you turned it. The elevator arrived and its doors opened on the right floor. The card reader accepted your tap. The traffic light cycled. If you flew somewhere, the aircraft's wings stayed attached and its engines kept running. If you or someone you love ended up in an emergency room, the lights didn't go out mid-procedure. None of these are achievements anyone will mention to you, because none of them are supposed to be achievements. They're supposed to be nothing: background facts about the world, as reliable as gravity.

They are not like gravity. Every "nothing" on that list is the output of a maintenance program, funded and staffed by specific people, running on a schedule somebody wrote down. This book has spent twenty-four chapters opening on an ordinary moment, a flush, a switch, a card tap, a blue dot on a map, and tracing outward through the delivery chain and the workforce that makes it work. This last chapter looks at a different moment: the one where something was supposed to fail, statistically, and didn't, because someone inspected it, replaced a part, or caught a problem early enough that you never heard about it. Reliability is not a property of concrete, copper wire, or steel. It's a job. This chapter is about the job itself, which is also, in a real sense, the argument this whole book has been making.

Three ways to fix something before it fixes you

Every physical system in this book eventually needs attention, and the attention comes in one of three basic forms, distinguished by timing.

Reactive maintenance, sometimes called "run to failure," means using a part until it breaks and then repairing or replacing it. For genuinely cheap components whose failure causes no cascading damage (a light bulb, a low-cost sensor), this is often the economically correct choice: there's no point inspecting a part more expensive to inspect than to replace. But for anything expensive, safety-critical, or connected to other systems that depend on it, reactive maintenance means the failure happens in public, usually at the worst possible moment, and usually costs more to fix than it would have cost to prevent.

Preventive maintenance replaces or services a component on a fixed schedule, based on its known failure characteristics, regardless of whether it shows any sign of trouble yet. A utility trims trees along a power line corridor every few years, not because that specific tree looks dangerous, but because vegetation growing into a transmission line is one of the most common causes of outages, and waiting for visible danger means waiting too long. An airline overhauls specific aircraft components after a fixed number of flight hours, whether or not that particular part has shown wear, because the manufacturer's data says components of that type start failing at a predictable rate past that point.

Predictive maintenance is the newest of the three: instead of a fixed calendar, it uses sensor data (vibration, temperature, oil chemistry, electrical signature) to monitor a component's actual condition continuously and flag it for service only when real evidence suggests it's approaching failure. A bearing that's about to fail typically vibrates differently, and heats up, well before it seizes; a transformer developing an internal fault changes the chemistry of its cooling oil in ways a sensor can detect months before a technician would notice anything by eye. Predictive maintenance can catch problems preventive maintenance's fixed schedule would miss (a part failing early) while also avoiding preventive maintenance's other cost: replacing parts that still had useful life left in them. Industry analyses from firms including Deloitte estimate that predictive maintenance programs can raise equipment uptime on the order of 10 to 20 percent and cut overall maintenance spending by roughly 5 to 25 percent depending on the industry, largely by eliminating both premature failures and unnecessarily early replacements.

Don't be confused: an inspection is not maintenance, and a maintenance plan is not funding. An inspection finds a problem. Maintenance fixes it. A written asset management plan describes what should happen and when. None of the three guarantees the others. A bridge can be inspected diligently, have its defects documented in writing for over a decade, and still collapse, because the inspection did its job and the repair budget didn't. Later in this chapter, the 2022 Fern Hollow Bridge collapse in Pittsburgh is exactly that failure mode: not a missed inspection, but an unfunded repair.

The reason this three-way split matters economically is that it isn't close. The U.S. Department of Energy's own operations and maintenance guidance for federal facility managers, drawing on decades of industry data, puts reactive repair at roughly three to five times the cost of the same job done as scheduled preventive work, once overtime labor, expedited parts shipping, and the secondary damage a failing part often causes to whatever it's connected to are all counted. A bearing that's serviced on schedule costs a part and an hour of a technician's planned time. The same bearing, left to seize on a Saturday night, can take down the machine around it, require emergency callout pay, and need a part flown in overnight. Well-run preventive programs are commonly credited with cutting total maintenance costs by something in the range of 12 to 18 percent compared to a reactive-only approach, a gap that compounds across an entire fleet of bridges, transformers, or aircraft components rather than a single part.

The thread running under every chapter in this book

None of this is a new idea introduced in this final chapter. It's the argument the whole book has been quietly building, one system at a time. The water utility running CCTV cameras through sewer mains to spot cracks before they become collapses, the electric grid trimming trees on a recurring cycle because untrimmed vegetation caused the 2003 Northeast blackout, the airline overhauling engine components on a fixed hours-based schedule and grounding any aircraft with an unresolved safety directive, the hospital load-testing its backup generators every month whether or not they've ever failed, the elevator undergoing an annual no-load safety test and a full-load test again every five years: these are not unrelated facts about unrelated industries. They are the same discipline, wearing a different uniform in every chapter.

That discipline has a name in engineering circles, even if it rarely gets one in casual conversation: asset management, sometimes narrowed to reliability engineering when the focus is a specific piece of equipment rather than an entire portfolio of infrastructure. The job is to decide, for every pipe, wire, beam, bearing, and generator a system depends on, what condition it's likely in, what it would cost to inspect versus replace it, and what order to act in when there isn't enough money or crew time to do everything at once. It is, at its core, a triage discipline applied not to patients but to physical objects, and like medical triage, it runs on data and probability rather than waiting to see what happens.

The people who decide what gets inspected, and when

The job title attached to that discipline is usually asset manager or reliability engineer, and it is a real, credentialed profession with its own professional body. The Society for Maintenance and Reliability Professionals (SMRP), chartered in 1992, certifies practitioners through the Certified Maintenance and Reliability Professional (CMRP) exam, the only credential of its kind accredited by the American National Standards Institute. The exam runs 110 questions across five areas SMRP calls its Body of Knowledge: business and management, equipment reliability, manufacturing process reliability, organizational leadership, and work management, and certification has to be renewed every three years with continuing education. The point of the credential is the same point this whole book has made about plumbers, wastewater operators, and pharmacists: deciding what gets inspected, and in what order, when there isn't enough money to do everything, is skilled work, not a matter of common sense applied by whoever happens to be on shift.

Beneath that credentialed layer sits an enormous, largely invisible workforce of inspectors: bridge inspectors trained and certified to federal standards, elevator inspectors licensed by the state or city they work in, aircraft maintenance technicians who sign off on every completed task in a logbook that follows the airplane for its entire service life, building facade inspectors working a city's five-year rotation. Almost all of them share one more thing in common: the timing of their work is deliberately invisible. Airlines schedule their heaviest maintenance checks, teardown-level inspections that can take weeks, during periods when an aircraft would otherwise be idle. Transit agencies run their most disruptive track work overnight and on weekends specifically because ridership is lowest then, testing rail fasteners, cables, and switches in the hours when the fewest people are inconvenienced by a train running slower or a station being closed. Utilities schedule substation and transformer maintenance for planned outage windows, announced in advance, rather than waiting for an unplanned one nobody scheduled. The public almost never sees this work happen, and that absence is not an accident. It's the goal. A maintenance program succeeding looks, from the outside, exactly like nothing happening at all.

A crack in an eyebar, and the law that followed

Almost none of this existed as a matter of federal law before a small number of very public disasters forced it into being, and the clearest example in the entire book is a bridge.

On the evening of December 15, 1967, at 4:58 p.m., during rush hour and holiday season traffic, the Silver Bridge, a suspension bridge carrying U.S. Route 35 across the Ohio River between Point Pleasant, West Virginia, and Kanauga, Ohio, collapsed without warning. Forty-six people died; two bodies were never recovered. Investigators traced the failure to a single small crack in one of the bridge's suspension chain eyebars, a structural element with no backup: the bridge's 1928 design had no redundancy built into that connection, so one hairline crack, invisible without specialized inspection, was enough to bring down the entire span once it propagated far enough.

The bridge had been standing for nearly forty years. It had presumably been looked at, informally, by various people over that time. What it had never had was a legally mandated, standardized, recurring safety inspection performed by a trained inspector looking specifically for the kind of defect that killed it. The Federal-Aid Highway Act of 1968, passed within a year of the collapse, directed the U.S. Secretary of Transportation to establish national inspection standards and a national inventory of bridges; the Federal-Aid Highway Act of 1970 required that every bridge on the federal-aid highway system actually be inspected under those standards; and the resulting regulation, the National Bridge Inspection Standards (NBIS), took effect in 1971, later expanded in 1978 to cover essentially every bridge longer than 20 feet on any public road in the country. The baseline requirement it set, inspection at least once every two years by a qualified, trained inspector, is the same rule most U.S. bridges are still inspected under today, more than half a century later. A 2022 update to the regulation added a risk-based option, allowing a well-maintained, low-risk bridge to be inspected on a longer cycle (up to 48 or, in some cases, 72 months) while a bridge already rated in poor condition still cannot legally go more than 24 months between inspections. The engineering logic runs directly backward from the disaster that created it: don't wait for a crack to be visible from the ground. Look for it on a schedule, whether or not anything looks wrong.

This pattern (a specific catastrophe producing a specific, lasting regulation) is not unique to bridges. It runs through nearly every safety system this book has described. New York City's facade inspection law, covered in the chapter on high-rise concrete, traces to a single death in 1979 and a partial collapse in 1998. The mandatory reliability standards that now govern the U.S. electric grid became legally enforceable only after the 2003 Northeast blackout exposed how toothless the old voluntary system actually was. Regulation, again and again in this book, has arrived after the fact, purchased with a body count, rather than installed in advance by engineers imagining the worst case. That's a genuinely uncomfortable thing to notice about how safety actually gets built, and this book has tried not to look away from it.

Turning inspection into law, at national scale

A rule requiring inspection is only half the system; the other half is requiring that inspection results turn into a funded plan of action, and that half arrived much later and much more quietly. MAP-21 (Moving Ahead for Progress in the 21st Century Act), passed by Congress in 2012, required every U.S. state transportation department to develop and maintain a formal, risk-based Transportation Asset Management Plan (TAMP) covering, at minimum, the bridges and pavement on the National Highway System. The plan has to show, in writing, how the state intends to reach and hold specific condition targets over time, not just react to whatever fails next. States that don't keep an approved plan face a real financial penalty: a mandatory diversion of highway formula funding away from general highway projects and into a narrower alternatives program until compliance is restored.

That may sound like paperwork, and in one sense it is. But it's the same shift this book has traced in system after system: from informal, individual judgment ("that bridge looks fine") to a documented, auditable, legally required process that assigns responsibility to a specific organization and gives regulators a lever to pull if the process isn't followed. It's the same shift that turned wastewater plant operation from a maintenance man's job into a state-licensed profession, and turned grid reliability from a gentleman's agreement between utilities into a FERC-enforced legal standard. Having a documented plan doesn't fix anything on its own. What it does is make the difference between an executed plan and a neglected one visible, on paper, to an auditor, a journalist, or a court, rather than leaving it as a private judgment call inside one maintenance department.

When the paperwork says fix it, and nobody does

Having the plan on paper and having the plan actually funded are not the same thing, and the clearest recent demonstration of that gap is not from 1967. It's from 2022.

Early on the morning of January 28, 2022, the Fern Hollow Bridge, carrying Forbes Avenue over a wooded ravine in Pittsburgh's Frick Park, collapsed into the ravine below, taking a city bus down with it. Ten people were injured; by good fortune, nobody died. The timing was almost too pointed to invent: President Biden was scheduled to arrive in Pittsburgh that same day to promote the newly passed federal infrastructure law. He visited the collapse site instead of the venue he'd originally planned.

The National Transportation Safety Board's investigation, released in February 2024, did not describe a hidden, unknowable defect. It described a documented one. Inspection reports had flagged holes and corrosion on the bridge's support legs as early as 2007. Reports from 2015 through 2021 repeatedly noted drains clogged with leaves and debris, letting water run down the bridge's legs instead of draining away, which stripped away the protective rust layer, called a patina, that would otherwise have slowed corrosion, and instead caused continuing, accelerating section loss in a fracture-critical member, a structural part with no redundant backup if it fails, exactly the same design vulnerability that doomed the Silver Bridge fifty-five years earlier. The NTSB's conclusion was blunt: maintenance and repair recommendations had been made repeatedly, in writing, for over a decade, and the City of Pittsburgh had not acted on them. The inspection system Congress built after 1967 had worked exactly as designed. It found the problem. What failed was everything after the inspection: the budget line, the work order, the follow-through. The bridge was rebuilt and reopened in under a year, largely funded by the same federal infrastructure law the president had come to promote, a new structure that will, eventually, get its own ribbon-cutting and its own clock of quiet, unglamorous inspections running underneath it.

What full repair would actually cost

Zoom out from any single bridge, and the scale of deferred maintenance across the country becomes a national accounting exercise, one the American Society of Civil Engineers has run every four years since 1998 in its Infrastructure Report Card. The 2025 edition gave the United States an overall grade of C, its highest ever recorded and an improvement from the C- given in 2021, driven partly by the federal infrastructure spending passed in 2021. Eighteen categories were graded individually, ranging from a B for ports down to a D for stormwater systems and public transit, and for the first time since the report card began, not a single category scored below a D.

The improvement comes with a number attached that undercuts any easy celebration. ASCE estimates that bringing all eighteen categories to a genuine state of good repair by 2033 would require $9.1 trillion in total investment; projecting current public and private spending trends forward over that same period yields roughly $5.4 trillion, leaving a gap of about $3.7 trillion, up from a $2.59 trillion gap estimated just four years earlier. ASCE frames the cost of not closing that gap in household terms: by its estimate, closing the investment gap would save the average American household on the order of $700 a year, mostly in costs that deferred maintenance currently pushes onto individuals: wasted time in traffic, vehicle repairs from poor road surfaces, and higher costs passed through from utilities absorbing their own deferred maintenance.

Bridges specifically, one of the more visible and heavily inspected categories, still hold their own C grade. Of the country's 623,218 bridges, roughly 6.8 percent, on the order of 42,000 to 46,000 structures depending on the year counted, are rated in poor condition (the category formerly labeled "structurally deficient"), and nearly half the entire national inventory sits in the "fair" category just above it, aging infrastructure that is not yet dangerous but is not being renewed as fast as it's deteriorating. ASCE's own bridge-focused analysis puts the ten-year funding gap for bringing every bridge in the country into good repair at roughly $373 billion. None of these are numbers a schedule can close by itself. They are numbers a budget has to close, year after year, against competing demands that are, almost by definition, more visible and more politically rewarding.

Ribbon cuttings, and the case for boring maintenance

That last point is the actual crux of why deferred maintenance keeps happening even in a country wealthy enough, in aggregate, to prevent it. Building a new bridge, a new terminal, a new hospital wing produces a ribbon-cutting: a date, a photograph, an elected official's name attached to something visibly new. Keeping an existing bridge properly maintained for the next thirty years produces nothing to photograph. It produces the continued absence of a collapse, which is not a event anyone can point a camera at. Public budgeting, especially at the state and municipal level where most of this book's infrastructure actually lives, systematically rewards the visible ribbon over the invisible prevention, because voters and officials alike respond to what they can see. Fern Hollow's own history makes the point uncomfortably well: the same week a president visited to celebrate new infrastructure spending, an old bridge that had been begging for repair money in writing since 2007 gave up entirely.

Two forces are pushing against that bias, unevenly. The legal one is the asset management mandate described above: a documented TAMP, reviewed by a federal regulator, creates a paper trail that makes chronic underinvestment visible to auditors even when it isn't visible to voters. The technical one is the growing maturity of predictive maintenance and the sensors that make it possible. Bridges are increasingly fitted with strain gauges and corrosion sensors that continuously report a structure's real condition rather than waiting for a technician's next scheduled visit; transformers are monitored for the same oil-chemistry changes described earlier in this chapter; aircraft engines stream vibration and temperature data during every flight, feeding predictive models that can flag a specific part for replacement before its scheduled overhaul date even arrives. None of this eliminates the underlying political problem, budgets are still set by people who respond to visible projects, but it does shrink the cost of the invisible option and sharpen the evidence for why it's worth funding anyway: a sensor reading is a harder thing to ignore in a budget hearing than a general appeal to caution.

Back to the day nothing happened

This book opened with a single sentence: modern life feels simple because extraordinarily complicated systems have been engineered to hide their complexity from the people using them. Twenty-four chapters have tested that sentence against a flush, a tap, a light switch, a supermarket shelf, a road, a concrete column, an elevator, an airplane seat, a delivered package, a card tap, a phone signal, a blue dot on a map, a hospital corridor at 3 a.m., a pharmacy counter, a 911 call. In every one of them, the simplicity was real at the surface and manufactured underneath, built by engineers, paid for by ratepayers and taxpayers, and operated by identifiable people doing skilled, often invisible work.

This chapter has tried to name the specific thread that ties all of those chapters together, because it's easy to miss if each system is looked at only once, in isolation. It is not electricity, or water, or concrete, or any single technology. It's maintenance: the continuous, funded, staffed, scheduled labor of deciding what will fail next, and doing something about it before it does. Every reliable system in this book depends on someone whose entire job is to notice decay before you do; on a budget line that has to be defended, every year, against something more photogenic; and on a regulation that, more often than any engineer would like to admit, exists because something like it failed once already, in public, at a cost measured in lives.

None of that will be visible the next time you flush, or flip a switch, or step into an elevator, or watch a blue dot move across a map on your phone. It isn't supposed to be. But it is there, running on a schedule, staffed by people with job titles you now know, funded (or not funded) by decisions made in budget meetings you'll never attend, every single day that nothing happens to you at all. That's not the absence of a story. It was the story, the whole time, and now you know where to look for it. 👉

The leap: what it replaced, and the work behind it

The disasters this book keeps returning to share an uncomfortable trait: most were not caused by a system nobody had built. They were caused by a system left to lapse. On August 1, 2007, at the evening rush, the I-35W bridge over the Mississippi River in Minneapolis dropped into the water, killing 13 people and injuring 145. The federal inspection regime that Silver Bridge created decades earlier had done its job of watching, the span had been rated structurally deficient since 1991, but the underlying flaw sat unaddressed for sixteen years while traffic and resurfacing loads grew heavier on top of it. The pattern recurs at every scale below the spectacular one. Roughly 240,000 water mains break each year in the United States, many of them pipes past their useful life that funding never reached in time. Deferred maintenance is not a dramatic act. It is the slow, boring choice to postpone, made one budget cycle at a time, until the postponement arrives all at once.

What continuous upkeep buys, against that, is almost impossible to photograph, but it is possible to count. The installation, maintenance, and repair workforce in the United States numbered about 4.8 million people in 2024, roughly 1.6 million of them general maintenance and repair workers alone, and the labor is worth more than the headcount suggests: the U.S. Department of Energy's own guidance puts scheduled preventive work at three to five times cheaper than the emergency version of the same job. Set that against the American Society of Civil Engineers' 2025 accounting, a C-grade infrastructure and a $3.7 trillion gap between what full repair would cost and what current spending will actually deliver, and the trade becomes legible. The gap is not a mystery about engineering. It is the sum of prevention deferred, priced out over a decade, waiting.

For you, the payoff is the ordinary day this book opened on. Somebody trimmed the tree before it fell across the line that feeds your block. Somebody load-tested the generator in the hospital where a relative is being operated on, months before anyone needed it. Somebody flagged the corroding beam, replaced the switch, cleared the drain that would otherwise have rotted a bridge leg from the inside. A single week of that work simply stopping would not look like a movie. It would look like a water main under your street letting go, then a transformer nobody inspected, then an elevator held for a test that never came, then a bridge downgraded to one lane, each failure small and local and, taken together, a quiet unraveling of the reliability you had stopped noticing. The absence of catastrophe is not luck; it is the accumulated result of people you will never meet doing unglamorous work slightly ahead of when it was strictly required.

Real-world examples and recent developments

The maintenance-versus-funding gap described above for the Silver Bridge and Fern Hollow is not a closed chapter of history; recent collapses, reopened investigations, and new inspection technology keep adding fresh, named cases to the same pattern.

  • Champlain Towers South (June 24, 2021): a 12-story beachfront condominium in Surfside, Florida, partially collapsed and killed 98 people, years after a 2018 engineering report from Morabito Consultants had already flagged "abundant" cracking and water damage around the pool deck and parking garage. A $15 million repair program had been approved before the collapse, but the structural work had not yet started, the same gap between a documented finding and a funded repair described above for Fern Hollow, just in a residential tower instead of a bridge.
  • Francis Scott Key Bridge (March 26, 2024): the Baltimore bridge collapsed after the container ship Dali lost power twice and struck a support pier, killing six highway workers on the span. The National Transportation Safety Board's final report traced the blackout to a single loose wire in the ship's electrical system, and separately found that the bridge itself had never been assessed for the risk posed by a vessel roughly ten times the size of the largest ship its 1977 design had accounted for, an inspection gap rather than a maintenance one.
  • Sunshine Skyway Bridge (May 9, 1980): forty-four years before the Key Bridge fell, the bulk carrier Summit Venture struck a support column of the original Tampa Bay span during a storm, killing 35 people and prompting the American Association of State Highway and Transportation Officials (AASHTO) to issue its first vulnerability guidance for bridges at risk from ship strikes. That guidance existed for more than four decades before Baltimore's Key Bridge failure showed how unevenly it had actually been applied.
  • State drone (UAS) bridge inspection programs: transportation departments in Minnesota, Nevada, Utah, and California now fly drones over bridges to capture close-up imagery and LiDAR data that once required a lane closure and a truck-mounted inspection platform, a direct extension of the sensor-based, predictive logic described earlier in this chapter into a task that used to be done entirely by eye from a bucket lift.

Recent developments

  • The NTSB's final report on the Key Bridge collapse, released November 18, 2025, concluded the Maryland Transportation Authority and many similarly situated bridge owners were "likely unaware" of the vessel collision risk their structures carried, and issued new recommendations on assessing that risk nationally.
  • AASHTO is drafting its first update since 2009 to the national bridge design specifications covering vessel collision protection, working with the Federal Highway Administration in direct response to the Key Bridge collapse.
  • NIST's investigation into the Champlain Towers South collapse, still open as of this writing, reported on September 9, 2025 that the failure most likely began in a pool deck slab-column connection rather than the tower itself, with a full set of technical reports and new recommendations for existing-building safety standards expected in 2026.

Glossary

Reactive maintenance. Repairing or replacing a component only after it fails, sometimes the right economic choice for cheap, non-critical parts, but the most expensive option for anything safety-critical or connected to other systems.

Preventive maintenance. Servicing or replacing a component on a fixed schedule based on its known failure pattern, regardless of whether it currently shows signs of wear.

Predictive maintenance. Using continuous sensor data, such as vibration, temperature, or oil chemistry, to detect a specific failure developing and schedule service based on actual condition rather than a fixed calendar.

Asset management. The discipline of deciding what condition a large portfolio of physical infrastructure is likely in, what it would cost to inspect or replace each part of it, and in what order to act given limited money and crew time.

Reliability engineering. The applied discipline of ensuring a specific piece of equipment or structure performs its intended function without failure over a defined period, closely related to but narrower than asset management.

Deferred maintenance. Known, documented repair or replacement work that has been postponed, usually for budget reasons, and continues accumulating cost and risk the longer it goes unaddressed.

National Bridge Inspection Standards (NBIS). The U.S. federal regulation, enacted in 1971 after the 1967 Silver Bridge collapse, requiring qualified inspectors to examine most public bridges on a legally mandated cycle, at minimum every two years for bridges not qualifying for a longer risk-based interval.

Fracture-critical member. A structural component with no redundant backup, such that its failure alone can cause the collapse of the entire structure it supports.

Transportation Asset Management Plan (TAMP). A formal, risk-based plan that federal law (since the 2012 MAP-21 act) requires every U.S. state transportation department to maintain for bridges and pavement on the National Highway System.

CMRP (Certified Maintenance and Reliability Professional). An ANSI-accredited certification, administered by the Society for Maintenance and Reliability Professionals, for practitioners in maintenance, reliability, and physical asset management.

Structurally deficient / poor condition. A federal bridge condition rating indicating significant deterioration in a bridge's deck, superstructure, or substructure, not an imminent collapse warning on its own but a signal that inspection and repair priority should rise.

Condition monitoring. The ongoing measurement of a machine's or structure's real-time physical state, using sensors, as the data source that makes predictive maintenance possible.

Planned outage. A deliberate, scheduled shutdown of a system or piece of equipment for maintenance, timed specifically to minimize public disruption, as opposed to an unplanned failure.

Investment gap. The dollar difference between what a full state of good repair for an infrastructure category would cost and what current funding trends are actually projected to spend on it.

Sources and notes

Open questions

  • The exact cost multiplier between reactive and preventive maintenance varies by industry and source, commonly cited between three and five times, and should be read as a well-supported range rather than a single precise figure.
  • ASCE's investment gap and total-need figures are recalculated every four years and depend on assumptions about future federal funding levels that were still politically unsettled (particularly around the 2026 expiration of current federal infrastructure authorizations) at the time of writing.
  • How much predictive maintenance and sensor-based condition monitoring will actually shift public infrastructure budgeting, as opposed to private industrial maintenance where adoption is further along, remains an open, active question rather than a settled trend.