How a disciplined battery qualification and replacement method returned two aging UPS units to verified service.
GDF completed this battery-replacement engagement for a Quebec industrial manufacturer, anonymised at the customer’s level of detail: no price, measured site value, or identifying configuration is published, and the anonymisation also withholds the exact UPS models and battery configuration. The engagement facts trace to an anonymised engagement attestation in the accompanying evidence ledger, and the engineering method is presented exactly as applied.
Operating context
The customer operates a production facility in Quebec where two single-phase, double-conversion uninterruptible power supply (UPS) units protect production-critical loads. Double-conversion units carry the load through their inverter continuously, so the battery system is the only ride-through energy source when the utility supply fails or sags. When the batteries in such a unit age past their useful life, the UPS chassis may remain healthy while the protection it exists to provide quietly disappears.
Like most commercial UPS in this power class, both units used strings of sealed valve-regulated lead-acid (VRLA) monobloc batteries. Commodity VRLA UPS batteries commonly deliver roughly three to five years of practical service life in float duty, and that life depends strongly on temperature, charge quality, and discharge duty. Neither unit had a documented battery replacement within that window.
Observed condition and business risk
The engagement began the way most battery projects begin: not with a failure, but with uncertainty. The batteries had aged beyond the typical VRLA float-service band, and the site had no current test evidence showing how much real capacity remained. That combination is the classic silent-risk profile. A UPS with exhausted batteries passes casual inspection, reports normal status on its front panel for most of the time, and then delivers seconds instead of minutes when an outage finally arrives.
The business risk was therefore framed honestly: not “the UPS is failing” but “the protection can no longer be demonstrated.” For production-critical loads, an undemonstrated ride-through is a risk decision by default, made silently every day the equipment stays in service. GDF’s role was to replace that silent default with a documented decision.
Engineering qualification and constraints
GDF does not treat battery age alone as proof of failure, and it does not treat a green status light as proof of health. Qualification for this engagement covered the elements that recognized stationary-battery practice expects before a replacement recommendation is made.
First, identity and configuration: the exact UPS models, their internal battery arrangement, and the manufacturer’s approved replacement configuration. A useful screening habit at this stage is dividing the nominal DC bus voltage by 12 to cross-check the expected count of series-connected 12 V monoblocs. That relationship is a screening check only. It applies to confirmed series strings of approved 12 V blocks and must never be extrapolated to other cell formats, centre-tapped arrangements, parallel strings, or integrated lithium cabinets; the manufacturer’s battery configuration documentation governs the replacement design.
Second, condition history: visual inspection for swelling, terminal corrosion and heat marks, the units’ alarm and event history, and the ambient conditions the strings had lived in. Third, the recognized end-of-life basis: stationary-battery practice, applied to VRLA strings through IEEE 1188, treats 80 percent of rated capacity as the replacement threshold, because capacity loss accelerates past that point and ride-through predictions stop being trustworthy.
The main constraints were operational: the loads were production-critical, so the work had to be planned around the plant’s operating schedule, and the replacement had to be completed within a single coordinated visit for both units.
Two environmental points from the qualification are worth passing on because they apply to most sites. First, temperature is the dominant life factor for VRLA batteries: sustained operation above the reference ambient of about 25 degrees Celsius shortens service life materially, so a string’s calendar age should be read against the room it lived in. Second, the timing of replacement is itself a risk decision. A planned replacement is scheduled around production, uses stock that has been verified in advance, and finishes with an orderly verification. An emergency replacement after a failed ride-through happens on the outage’s schedule instead, often at premium cost and with whatever batteries can be sourced quickly. Aging strings reliably convert the first situation into the second when left long enough.
Options considered and decision basis
Three options were considered, which is the normal decision set for aged strings on otherwise healthy UPS hardware.
Continue in monitored service. Rejected: with the strings beyond the typical service band and no current capacity evidence, continued service would have preserved the silent-risk profile rather than resolving it.
Test first, then decide. A legitimate route when age is ambiguous or the strings are young enough that real capacity may remain. Here, the combination of string age and the absence of any baseline made a test program an added cost on the path to the same likely outcome. The customer elected to move directly to replacement, a decision the age evidence supported.
Planned replacement. Selected. Replacement restores the protective function, resets the service clock, and creates the documented baseline that future condition assessment needs.
With replacement selected, the selection question became which battery. GDF’s preferred replacement cell for this duty is a high-rate, flame-retardant VRLA monobloc with a UL 94 V-0 case. Preferred does not mean automatic: the specific cell was confirmed against the units’ charger profile, discharge duty, physical dimensions, terminal arrangement, and warranty conditions before being ordered. The flame-retardant casing matters because UL 94 V-0 is the strongest of the common vertical self-extinguishing flammability classifications for enclosure plastics, a meaningful property for batteries that live inside occupied facilities rather than dedicated battery rooms. GDF maintains a dedicated guide to flame-retardant UPS batteries and where the V-0 casing class matters.
Scope of work
The agreed scope covered both UPS units in one coordinated site visit: supply of the verified replacement monoblocs, removal of the aged strings, installation and connection of the new strings, functional verification of both units, and removal of the old batteries for recycling through the appropriate stream. Lead-acid batteries are among the most successfully recycled industrial products, and returning them to that stream is part of a professional replacement rather than an optional extra.
The scope deliberately excluded items that were not supported by evidence or need: no electrical modification of the UPS units, no runtime upgrade beyond the approved configuration, and no monitoring retrofit. Keeping replacement scope tight is itself a discipline; a battery visit is not the moment to improvise system changes.
Implementation and verification
The replacement followed the method in Figure 1. Work on each unit began with isolation and lockout appropriate to the equipment and the site’s rules, because a UPS battery string is a stored-energy source that does not de-energize just because the utility feed is off. The aged blocks were removed, the replacement blocks installed, and every connection made with torque control and polarity verification, followed by a thermal screen of the completed connections.
A practical note for readers planning similar work: replacing batteries on an in-service UPS temporarily reduces or removes the ride-through protection while strings are disconnected. That exposure window cannot be fully eliminated on a single-string system, so it is scheduled deliberately with the customer, which is exactly what the coordinated visit was for.
Verification is what turns a parts swap into an engineering deliverable. Each unit received a functional transfer test, confirming that the UPS carried its load correctly on the new strings and returned to normal operation cleanly. Both units completed the test and were returned to service. Following the replacement, GDF’s technician verified and reported in writing that both UPS units were fully functional, which was the engagement’s defined success condition.
The customer received the technician’s written report stating the verified outcome and follow-up recommendations, including one additional attention item; that written record establishes the baseline the next condition assessment will measure against.
Results, measurement basis, and limitations
The published result is deliberately precise about what was and was not measured. What is claimed: two UPS units received complete replacement strings of OEM-compatible, high-rate, flame-retardant VRLA monoblocs; both units were verified fully functional after the work; both returned to service with their protective function restored and documented in the technician’s written report.
What is not claimed: no runtime figure, in minutes, is published for the site, because no approved discharge test formed part of the published evidence, and runtime claims require a documented discharge basis with method, load state, and conditions recorded. No capacity percentage is published for the removed strings for the same reason. This restraint is intentional. A case study that invents the numbers its evidence does not contain teaches readers to trust the wrong things.
The limitation cuts both ways and is worth stating plainly: a functional transfer test demonstrates that the UPS and its new strings work together correctly under real load; it does not measure total stored energy. Sites that need a demonstrated autonomy figure should plan a controlled discharge or runtime test as a separate, deliberate step.
Practical lessons
Four lessons from this engagement generalize to most VRLA-based UPS fleets.
Age without evidence is a decision already being made. Once strings pass the typical service band with no capacity evidence, doing nothing is itself a risk acceptance, and it is being accepted silently. Naming that explicitly is often what unlocks a sound decision.
Selection is verification, not preference. A preferred battery line, however well chosen, earns its place on a specific site only after charger profile, duty, dimensions, terminals, and warranty conditions are verified against the manufacturer’s documentation.
The casing class is part of the engineering. Where batteries live inside occupied facilities, specifying a UL 94 V-0 flame-retardant case is a low-cost, high-value safety property, not a cosmetic option.
Verification defines completion. The engagement was complete when both units passed a functional transfer test and the result was documented, not when the last block was torqued down. That verification record is also the first data point of the new strings’ service life.
Sources and next step
Bracketed codes such as identify entries in the source and claims ledger accompanying this document: The ledger draws on GDF’s anonymised field case library, GDF’s internal battery selection and testing references, Schneider Electric’s published battery-system design guidance, UL’s published flammability classifications, and IEEE stationary-battery practice.
If your UPS batteries have aged past their service band, or their real condition cannot be demonstrated, GDF’s UPS battery repair and replacement service page describes scope, brands covered, and how to request an assessment. GDF Technologies services APC, Eaton, Vertiv, Liebert, Mitsubishi, Tripp Lite and other major UPS brands across Canada as an independent specialist.
GDF Technologies services every UPS brand covered here, across Canada. See our preventive UPS maintenance plans or call (514) 252-8324 for on site support.