Twelve years as an academic in power systems at the University of
Canterbury. My PhD was on interharmonic cross modulation in HVDC links.
Then twenty years in the
New Zealand electricity sector: a couple of years consulting to the
Electricity Commission first, and the rest at the Electricity Authority,
where I work across compliance, monitoring and policy.
The paper below carries the Authority’s
affiliation. The views in it are mine, not theirs. Everything else here
is my own time.
A wise man proportions his belief to the evidence.
David Hume, 1748 — no relation established
Twenty-nine years of half-hourly metering at New Zealand’s
grid exit points, measuring the reactive character of a whole power
system. It crosses from lagging to leading around 2016, and most of the
drift is on the demand side.
The likely mechanism is the mandated EMC filter inside every
switched-mode power supply, though that is a hypothesis and the paper
labels it as one. The inputs are records any system operator already
holds, so the measurement should travel.
Power factor across all 29 New Zealand electricity distribution
businesses, from public EMI data. Sixteen years of it, plotted so you
can watch the reactive power drift network by network.
An independent physics analysis of the Langtang Lirung collapse of
26 August 2026, and the flood it sent down the Lhende Khola, the Bhote
Koshi and the Trishuli. Reported elsewhere as the Bhote Koshi flood. As
at 8 September, 1,342 people were confirmed dead in Nepal and 4,886
missing, with 43 dead and 519 missing on the Chinese side. Every figure
is provisional.
One narrow question: where did the water actually come from? I think
it was largely already in the river. Melting ice is expensive and
falling is cheap, so the fall cannot supply the ~20 Mm³
officially recorded as excess. That part is an energy budget, not a
model.
How big the collapse itself was is a harder question and I have not
answered it. The early estimate is under revision and looks too small,
seven versions of the routing model have failed the tests I set them,
and it may end up withdrawn rather than resolved.
Scar · 4,928 mDevghat · 186 m
Channel long section, 199 km, from the project’s own
routing data.
Still going, so treat the size numbers as
working numbers rather than results. Five findings retracted in the
first week and more corrected since, including one where I quoted a
figure with a precision the source never claimed. All of them still up,
with the reasons and the dates.
Rogaines and adventure racing these
days, mostly in the South Island. Before that a lot of caving, quite a bit
of it genuine exploration in the deep parts of New Zealand’s biggest
systems, where the map runs out. Cave search and rescue too. Pipi the dog
comes to the training but not the races.
