Browse
Log in / Register
Share
Cook Nuclear Plant in Bridgman, Michigan, has filed two defect notices with the U.S. Nuclear Regulatory Commission this month: one for a latent flaw that became an internal short inside a Unit 1 safety battery cell, another for four failed starting-air compressors that support the emergency diesel generators. The battery defect was reported September 17 after a routine quarterly test; the compressor notice came the next day. As of September 25, 2026, Cook runs weekly voltage checks on the battery string and is installing older-design replacement compressors.
Anyone who has nursed a dying car battery through a Michigan winter knows the pattern. The starter turns over a little slower each morning, and eventually one cell inside the case gives out and drags the whole battery down with it. The safety batteries that back up a nuclear plant can fail the same way, and Cook just saw it happen.
Cell 50 of 1-BATT-CD, one of Unit 1's safety-related station batteries, dropped below the voltage its technical specifications require during a quarterly test. Crews tried a full recharge, but the cell would not hold it, and workers pulled the cell from the string about a week later.
Manufacturer C&D Technologies took the cell apart and found an internal short. A clump of lead-oxide paste had migrated through the separator and contacted the adjacent negative plate, leaving the cell unable to maintain stable current while charging. No amount of charging was ever going to bring it back.
That may sound like an obscure chemistry failure, but it's a classic lead-acid failure mode. A cell is essentially a stack of plates coated in lead-oxide paste, with thin separators keeping the positive and negative plates apart. One displaced blob of paste bridging the gap is all it takes to produce a hard short inside a sealed container. This particular blob had been sitting quietly in the box since the cell was built in February 2016.
The battery is not there to generate power. It sits in reserve for the worst-case moment: if Cook loses its offsite power lines and, at the same time, its emergency diesel generators — the station blackout scenario — the battery keeps instruments, controls, and other essential equipment running while crews restore power. It is the kind of event that defined the 2011 Fukushima accident.
A station battery is not a monolithic block. It is a long string of cells wired in series, and the string is only as good as its weakest cell. The filing does not say how many cells 1-BATT-CD contains or exactly how far the voltage dropped, but it does spell out the worst case: if the same flaw exists in several cells at once, the battery could fall short on capacity, or on the run time it must deliver during a blackout.
Cook's filing notes the defect gives no visible warning until the short actually happens. There is no way to spot it by looking. The practical defense is monitoring, which is why the plant now measures the battery's terminal string voltage every week instead of waiting for the quarterly check that caught this one.
Weekly voltage readings are a faster tripwire, not a cure. Real capacity loss is still only confirmed by exercising a battery under load, which is why periodic discharge testing remains the core of IEEE 450, the industry's maintenance standard for vented lead-acid batteries. But cutting the interval between checks means a cell that starts dragging its string down is much more likely to be found within days, not months.
The battery was not the only piece of backup equipment generating NRC paperwork this month. On September 18, Fort Worth-based Paragon Energy Solutions filed its own defect notice after Cook reported that four starting-air compressors for the emergency diesel generators had failed.
Those diesels do not start with a key. They need a strong burst of compressed air to spin up, and the compressors in question are the machines that keep the air storage tanks topped off. When they cannot hold pressure, the diesels are left to run on whatever air is already in the tanks — a margin nobody wants to rely on.
Paragon's teardown traced the problem to a gasket. On the failed units, the gasket on a second-stage valve plate had broken apart, and fragments lodged in the valve seat. The resulting back-pressure lifted the first-stage relief valve and pulled discharge pressure below what the diesels need. All four compressors — parts H7100X10-MOD and H7100X15 — were built in 2022 or later with an updated valve plate design.
Paragon's remedy is a deliberate step backward. The company is shipping Cook replacement compressors built with the original valve plate design, and it has asked the plant to keep a close watch on the units still in service while the root-cause investigation with Ingersoll Rand continues. A redesigned part losing to the part it replaced is not a flattering outcome, but a gasket that holds is worth more than a design that does not.
Put it in context: when Europe's largest nuclear plant, Zaporizhzhia in Ukraine, lost its last external power line, emergency diesel generators were the only thing keeping the shut-down reactors cool, and crews stretched fuel and air for more than two weeks. Diesels that start are everything, and starting air is the quiet part of that game.
Neither failure moved the reactor's calendar. Unit 1 shut down at 3 a.m. on Saturday, September 19, for its planned refueling outage, and Unit 2 keeps running at 100 percent power. Indiana Michigan Power said customers would not notice any difference. The failed battery cell had already been replaced about a week after the test flagged it, before the outage began.
Both notices were filed under 10 CFR Part 21, the federal rule requiring utilities and suppliers to report defects in safety-related components. It acts as a recall system for nuclear hardware and catches impressively small things. Earlier this month, the same trail surfaced acrylic paint inside a breaker at another plant. This time it was a displaced clump of paste and a fractured valve gasket. There is no sign either defect endangered Cook; both were caught by scheduled testing and put on the public record. That is the system doing its job.
Cook owes the NRC a full written report on the battery within 30 days of the September 17 notice, so a submission is due around mid-October. Paragon still has open root-cause work with Ingersoll Rand. Replacement compressors built on the original valve plate design are on their way to the plant. And the terminal voltage on 1-BATT-CD is now read every week.
The larger question is whether this ends at Cook. The failed cell was a C&D LCR-33 NUC, a lead-acid model qualified for nuclear service and built in February 2016. A latent defect that waited more than a decade to surface raises an obvious common-cause question: what about the other cells in that same string, and what about cells from the same production era installed at other nuclear stations? Part 21 duties apply to suppliers, not just utilities, so if C&D determines the flaw could affect other cells, the NRC's public docket could see follow-up notices in the coming months. Until then, Cook's answer is the unglamorous one that tends to work best: measure the system more often, and catch the next weak cell long before it matters.
