Pull to refresh
Logo
Battery self-tests synced to the minute cause grid frequency dips

Battery self-tests synced to the minute cause grid frequency dips

Built World

A Jülich-led team traced a regular 60-second wobble across power grids on five continents to battery software that checks in at the top of every minute

Overview

Updated 1 hour ago

For years, engineers at the transmission operator TenneT watched the frequency of the European grid dip, briefly, every 60 seconds. The drops were too small to threaten supply. They were regular enough to be strange.

A team led by Dirk Witthaut at Forschungszentrum Jülich traced the pattern to battery backup systems that check their charge at the top of every minute. Network clocks make the checks fire together, so many units draw power at the same instant. Spreading the checks out would remove the spike.

Why it matters

Every battery backup that self-tests on the minute adds to a grid wobble that wears on turbines. Randomizing the timing costs little.

Questions about this story

Free account needed to ask — your question is kept and asked for you right after sign-up. Answers are public.

No questions yet — be the first to ask.

Key Indicators

60 s
Period of the grid wobble
Frequency records show a narrow peak at exactly one minute, about 0.0167 hertz.
12
Electricity systems measured
Frequency records from grids on five continents showed the same one-minute peak.
80
Metering points on the Jülich campus
Building-level measurements traced the pattern to battery-backed uninterruptible power supplies.
20%
Irish electricity demand from data centres with battery storage
Ireland's one-minute signal grew fastest as these installations expanded.
4×
Growth of Ireland's one-minute signal, 2016 to 2021
The amplitude of the signal quadrupled in five years.

Voices

Curated perspectives — historical figures and your fellow readers.

Sign up to generate historical perspectives on this story.

People Involved

Organizations Involved

Timeline

2021 October 2026

2 events Latest: Yesterday
  1. Study pins the wobble on synchronized battery checks

    Latest Research

    A Jülich-led team reports in Nature Energy that battery systems querying their charge at the top of each minute produce the recurring dip, and calls for randomized timing.

  2. Ireland's one-minute signal quadruples

    Measurement

    Between 2016 and 2021, the amplitude of the one-minute frequency signal in Ireland grew fourfold as data-centre battery storage expanded.

Scenarios

1

Battery makers randomize self-test timing

Possible Resolves by End of 2027

Discussed by: Dirk Witthaut and co-authors, who call for randomized scheduling in Nature Energy

Witthaut's group says the fix is to spread the checks randomly rather than start them all at the same moment. A software update to one battery block in the study reduced the strict periodicity but left some synchronization in place, so a full fix depends on vendors changing defaults across product lines, not on one patch.

2

Standards bodies set a rule for self-test timing

Uncertain Resolves by Q2 2028

Discussed by: Grid engineering analysts at gridperspective.com, who note the study's call for adaptive timing offsets

No binding standard governs these battery checks, which is why manufacturers independently settled on 60 seconds and the start of the minute. A requirement from an EU grid-code body or an international standards organization could mandate offsets or randomization. Slower than vendor action, but harder to reverse.

3

Storage growth outpaces any fix

Possible Resolves by End of 2028

Discussed by: Witthaut and co-authors, who warn the problem could grow with storage expansion

If defaults stay and nothing is standardized, each new battery installation adds to the synchronized load step. Ireland's experience is the template: the signal quadrupled in five years as data-centre batteries multiplied. The same trajectory in other markets would mean more wear on hydro turbines and other equipment.

Historical Context

2 moments from history that rhyme with this story — and how they unfolded.

December 1999 - January 2000

Y2K date rollover (1999-2000)

As the calendar turned to 2000, computer systems worldwide were scheduled to evaluate dates written with two-digit years. Companies and governments spent years auditing code and running tests, because a shared clock would make many independent systems act at once.

Then

Few serious failures materialized, and the rollover passed with scattered, limited disruptions.

Now

The episode made clear how much infrastructure depends on synchronized time, and it established practices for testing clock-driven software before a deadline.

Why this matters now

Both cases turn on the same mechanism: a shared clock causing many separate machines to act simultaneously, with consequences that only appear at the system level.

June 2012

Leap second outage (2012)

An extra second was added to Coordinated Universal Time on 30 June 2012. A flaw in how some Linux systems handled the adjustment caused servers to hang, and several high-traffic websites reported outages.

Then

Service disruptions lasted hours at some sites; vendors issued patches.

Now

Companies began auditing how their software handled clock adjustments, and the leap second stayed contentious for years.

Why this matters now

A one-second change to a shared reference clock knocked out systems that were individually well built. The 60-second grid wobble comes from the same kind of coupling.

Sources

(5)