30 October 2026 · By Oleg Gavrilov

Some seals may cost only a few dollars. The equipment that depends on them can be worth millions. That gap between the price of the part and the consequences of getting it wrong is why we are building Seallect.

On 28 January 1986, the Space Shuttle Challenger broke apart 73 seconds after launch, killing all seven crew members. Rogers Commission, Chapter III. The Rogers Commission traced the accident to a failed pressure seal in the right solid rocket motor. The joint used O-rings, whose response was impaired by the cold. Rogers Commission, Chapter IV. But the investigation went beyond the rings themselves: it found a faulty joint design and failures in the decisions that allowed the launch to proceed. Rogers Commission, Chapter V.

Few sealing failures have consequences on that scale. Most are quieter: a leaking cylinder, a pump taken out of service, a repair that has to be done twice. Yet the same question runs through them. Will this sealing solution work in this equipment, under these conditions?

Answering that question is the business we want Seallect to be in.

Think of a hydraulic cylinder lifting a heavy load. Its seals must contain pressurised oil while a rod moves in and out. In a pump, a sealing system must control leakage around a rotating shaft. In a valve, one seal may keep fluid from escaping outside, while another closes off the flow inside.

The job changes with the machine. So does the solution. Trelleborg’s industrial sealing range illustrates the variety:

Some are simple standard parts. Others are engineered assemblies made from several materials, with supporting components and precise installation requirements. Their purchase prices vary widely. What they share is a demanding role at the boundary between pressure, movement, fluids and the outside world.

Choosing correctly means understanding the geometry, materials, pressure, temperature, motion and fluid compatibility together. A part that looks right and fits the nominal dimensions can still be wrong for the job.

That is the wider field Seallect is being built to address: industrial sealing solutions across cylinders, pumps, valves and rotating equipment.

Fortune Business Insights estimates the global industrial seals market at US$15.61 billion in 2025. Its forecast projects US$26.95 billion by 2034, with a compound annual growth rate of 6.5% over its 2026–2034 forecast period. Read the market estimate and forecast.

These commercial estimates give a sense of scale. They do not tell us how much is lost when seals fail, or what share of the market Seallect could serve.

To understand the consequences of a wrong choice, look at the equipment around the seal.

A leaking seal can mean much more than replacing a component. The machine may need to be shut down, emptied, dismantled, repaired and tested before production can restart.

In a process plant, a leak can also become an emissions or safety problem.

Siemens’ “The True Cost of Downtime 2024” estimated that unplanned downtime costs the Fortune Global 500 almost US$1.4 trillion a year, equivalent to 11% of revenue. That is an extrapolation from research based on 181 online interviews covering April 2019 to March 2023.

Large plants in the surveyed sectors lost an average of 27 hours a month. Estimated hourly losses ranged from US$36,000 in fast-moving consumer goods to US$2.3 million at a large automotive plant. Siemens report, pp. 2–4 and 15.

Those figures cover unplanned downtime across causes, not losses attributed specifically to seals. But they explain why a maintenance team urgently trying to restart a line is concerned with far more than the price of a replacement part.

The US Environmental Protection Agency’s 2007 leak detection and repair guide cited estimates of about 70,000 tons of volatile organic compounds and about 9,000 tons of hazardous air pollutants released annually from equipment leaks at US refineries and chemical manufacturing facilities.

Earlier studies cited in the guide attributed more than 90% of equipment-leak emissions to valves and connectors. These are historical estimates, but they show why sealing and maintenance deserve attention. EPA’s 2007 guide, pp. 2–3.

Seal-system design can also change emissions during normal operation. An EPA guide from 2006 reports 40–200 standard cubic feet per minute for a two-seal wet system on a natural-gas centrifugal compressor, including gas released when the seal oil is degassed.

For two dry seals, it gives 1–6 standard cubic feet per minute. This is a comparison of systems, not just leakage across their seal faces; conversion suitability and economics depend on the compressor. EPA’s October 2006 guide, pp. 1–2.

In critical equipment, the consequences extend beyond lost production. Challenger remains an extreme reminder that a seal cannot be considered in isolation from its joint design, operating conditions or the engineering decisions around it.

Better selection is part of that responsibility, alongside installation, maintenance and sound judgement.

The everyday examples are less dramatic, but familiar to anyone in sealing work.

In one hydraulic cylinder repair, a guide ring was about 3.2 mm thick radially where about 2.5 mm was needed. The rod would not go through. The guide ring was not itself a seal, but that one mismatch made the repair kit unusable.

In another case, a production line was already stopped when a replacement sealing profile was checked against the press. The profile was 12.3 mm high. The groove measured 10.5 mm. A difference of 1.8 mm was enough to complicate an urgent repair.

A third case involved a proposed change to an O-ring’s cord diameter. Once the actual groove dimensions were confirmed, the supplier calculated that the change would reduce compression from about 18% to about 7%, below its recommendation for that particular assembly. The apparently simple correction needed to be reconsidered.

These cases come from our own sealing work. They show how a plausible part or dimension can reach a specification before the full picture has been checked.

After more than 20 years in sealing sales, I have seen how much work can sit behind a seemingly simple recommendation.

Four problems keep recurring.

The data often exists, but across catalogues, drawings, compatibility charts and certificates. Manufacturers use their own compound codes and document formats. Even where sizes follow recognised standards, tolerances, material properties and test conditions still need to be compared. James Walker alone offers O-rings in more than 100 elastomer grades, an indication of the choices within just one product family. James Walker’s O-ring range.

A matching polymer name or nominal size is a starting point, not proof of equivalence. Manufacturer tools can be useful, but an engineer comparing several suppliers may still have to reconcile their information by hand.

A certificate may apply to a particular compound, product or use. A broad claim such as “suitable for food applications” does not replace checking the relevant document and its limitations. The question is whether the evidence covers this product in this application.

A drawing may belong to an earlier machine revision. An old seal may be worn or distorted. An estimated dimension may lose its warning label as it passes from one person to another. By the time an order is placed, an assumption can look like a confirmed fact.

Experienced people already know to make these checks.

The challenge is bringing the information together reliably, especially when a machine is down and time is short.

Seallect is in development. Our ambition is a brand-neutral platform that helps people define the sealing solution their equipment needs, then compare the manufacturers that can supply it.

The starting point would be the application: what moves, which fluids are present, the pressure and temperature range, the available space and any required approvals. The technical requirements would come before the brand or part number.

AI could make that process easier to navigate. It could help users describe a problem, find relevant information and understand the trade-offs.

But a fluent answer is not enough to support an engineering recommendation. The data and the checks underneath it matter more.

Our planned approach has three parts:

The potential value is practical: fewer specification errors reaching the workshop, less time spent cross-checking documents, clearer comparisons between suppliers and a record of why a solution was chosen. Those are outcomes we still need to test and measure with users.

We are starting with O-rings and back-up rings so we can build and test the approach within a defined scope. That is the first step, not the boundary of the platform.

The wider plan includes hydraulic and pneumatic sealing, pump and valve applications, and rotary sealing systems. Each expansion will need suitable data, application-specific checks and engineering review.

Software will not measure a groove, establish which drawing matches the actual machine or replace a qualified engineer’s judgement in a critical application. It can help make the information easier to check before a decision becomes an order.

The seal may be a small part. The decision deserves better information.

That is why we are building Seallect.

Back to Journal · Home