ICHQP 2026 — Attendee Guide

Why this appendix exists. Optional background, parked out of the way of the conference plan. Written to be read on the plane, in order — not required to use the schedule above. If you absorb only the two mental models in A.1 and the vocabulary in A.3, you will follow essentially every talk at ICHQP. Everything after that is depth.

Anything below marked (contested) is genuinely argued about by the people in the room. Don't state it as settled fact at a coffee break.

Model 1 — Power quality is a compatibility problem, not a purity problem.

The instinct of every newcomer is that the goal is a perfect 50 Hz sine wave. It is not, and it never was. Perfect power is infinitely expensive and nobody needs it. The actual question the field asks is a three-way negotiation:

Set emission limits too loose and equipment breaks. Set them too tight and nothing can connect and the energy transition stalls. Set immunity too high and every appliance costs more. The entire field — every standard, every product, every argument you will hear this week — lives in the gap between those three numbers. Utilities lobby from the emission side. Manufacturers lobby from the immunity side. The standards bodies are the referee. That's the game.

Practical tell: when someone says a number like "8% THD", your first question should always be which of the three levels is that? — and the follow-up, at which point in the network? The same number means completely different things at a customer's terminals versus at the substation.

Model 2 — One trend is breaking the field: everything is becoming a power-electronic converter.

The 20th-century grid had a clean division of labour. Big spinning machines — coal, gas, hydro, nuclear — made the voltage. Their physical inertia and their windings defined the frequency and the impedance. Loads were mostly passive: motors, heaters, incandescent bulbs. They drew current in roughly the same shape as the voltage handed to them.

Both halves of that have now inverted:

• Loads became switching converters. LED lighting, variable-frequency drives, EV chargers, data-centre power supplies, heat pumps, induction hobs, PV inverters in reverse. These don't draw a smooth sine of current — they chop it, at kilohertz, and the chopping shows up on the grid as distortion.
• Sources became converters too. Solar and wind and batteries connect through inverters. An inverter has no physical inertia and no fixed impedance — it has a control algorithm, and its apparent impedance is whatever that firmware decides, differing by frequency and updated by software release.

That second point is the deep one, and it's the reason this conference exists in its current form. Grid stability used to be governed by physics you could measure once and write down. It is increasingly governed by proprietary control software from a dozen vendors, interacting in ways nobody simulated, in a grid where the stabilising spinning mass is being retired. When two vendors' control loops fight each other at 3.5 Hz across a national grid, that is a power quality event with no 20th-century analogue.

Run that lens over the 2026 programme and every hot topic falls out of it as a consequence: grid-forming inverters (can software fake inertia?), supraharmonics (the switching frequencies themselves, 2–150 kHz), sub-synchronous oscillations (control loops fighting), DC power quality (if everything's a converter anyway, why keep converting to AC?), EV charging and data centres (the two biggest new converter loads on earth). Not five topics. One topic, five faces.

The IEC 61000 numbering system, decoded. The middle digit is the whole trick:

• 61000-1-x — general: definitions, rationale, methodology
• 61000-2-x — the environment. Compatibility levels: how much disturbance we agree to assume exists. (-2-2 public LV, -2-4 industrial, -2-12 public MV)
• 61000-3-x — limits. What a device or an installation may emit. (-3-2, -3-12 for equipment; the -3-6 and -3-7 Technical Reports for allocating a budget among large connecting customers)
• 61000-4-x — testing and measurement techniques. How you prove any of the above. (-4-7 harmonics, -4-15 flickermeter, -4-30 the PQ measurement bible)
• 61000-5-x — installation and mitigation guidelines · -6-x — generic immunity/emission

Read left to right and it's the compatibility model from A.1, made operational: -2-x says what the environment tolerates → -3-6/-3-7 divides that budget among the big connecting customers → -3-2/-3-12 constrains what manufacturers ship → -4-x says how you measure to settle the argument.

If you learn one "unsolved problem" well enough to discuss it, make it this one. It comes up in the keynotes, in Panel 4, in at least three technical sessions and on the exhibition floor, and understanding it makes you legible as someone who's actually paying attention.

How the gap got there. Two communities regulated the frequency spectrum from opposite ends and didn't meet in the middle:

• Below ~2 kHz — the power engineers. Well covered: IEEE 519, EN 50160, IEC 61000-2-x and -3-x. Mature, enforced, litigated.
• Above 150 kHz — the radio and broadcast community, protecting their spectrum via the CISPR standards. Strict, old, well-policed.
• 2 kHz to 150 kHz — nobody's. A regulatory no-man's-land.

Then modern power electronics moved in. This band is exactly where converters switch. And here's the genuinely perverse bit, documented in the literature: because the <2 kHz limits were enforced, product designers had an incentive to meet them by pushing emissions up into the unregulated band instead. The regulation didn't remove the energy; it relocated it somewhere nobody was measuring.

What it breaks in practice: EV chargers malfunctioning near each other, interference with utility ripple-control and powerline-communication signalling, audible noise from transformers and appliances, LED drivers behaving strangely, and a general category of "unexplained equipment misbehaviour" that gets written off because nobody's instrument can see the cause.

Why it stays unsolved — four distinct holes, not one:

1. No compatibility levels. Nobody has agreed how much of this the environment should be assumed to contain.
2. No mandatory emission limits. What exists is provisional (CISPR 15, for lighting) or merely recommended (IEC TS 62578, for active infeed converters). There's an immunity standard — IEC 61000-4-19 — but immunity without an emission limit is half a compatibility model.
3. Two incompatible measurement methods. IEC 61000-4-7 covers 2–9 kHz one way; 61000-4-30's annex covers 9–150 kHz another way, with different filtering and grouping. Results across the full range aren't directly comparable — and until 2025 the high-frequency annex was informative, i.e. not binding.
4. Impedance isn't specified. A voltage-based measurement is only reproducible if you know the grid impedance at that frequency, and the measurement standards don't tell you how to establish it. This is why you'll hear "we have no standardised impedance measurement above 9 kHz" repeated all week like a refrain — it's the blocker underneath the blocker.

Why a PM should care more than an engineer. A standards vacuum with real customer pain is the most valuable position in an instrumentation market. Most installed monitors physically cannot see this band — not just software, but sampling rate and, critically, the current and voltage transducers in front of the ADC. When the measurement method converges, there is a hardware replacement cycle waiting behind it. And the vendor whose method is credible, repeatable and public when that happens tends to become the reference. Watch which exhibitors can answer question 2 from the exhibition checklist without flinching.

(a) Voltage sags are the expensive phenomenon; harmonics are the famous one.

The conference is named after harmonics and the majority of the papers are about them. But harmonics mostly cost you slow, diffuse, hard-to-attribute losses — extra heating, shortened equipment life, the occasional capacitor bank. Voltage sags cost money in a single visible event: a 100-millisecond dip nobody notices in an office drops a semiconductor fab, a plastics line or a paper mill and destroys a batch. Note the mismatch — academic attention is on harmonics, industrial willingness-to-pay is on sags and interruptions. Mismatches like that are where products come from.

(b) The industry's headline cost figures are alarmingly old — and that is itself a finding.

Nearly every PQ pitch deck you will ever see rests on two studies:

• EPRI/Primen (2001) — power disturbances cost the US economy over $119 billion a year, of which roughly $15–24 billion is power-quality rather than outright outages. Survey-based, self-reported, 25 years old.
• Leonardo Energy / European Copper Institute (2008) — poor power quality costs EU business over €150 billion a year, about 90% of it industrial, with voltage dips and short interruptions responsible for more than half of all losses. Based on 62 participants across 8 countries, extrapolated to a continent.

Both are weak evidence carrying enormous commercial weight. There is no modern, credible, bottom-up cost-of-power-quality study. If you want a conversation-starter that makes senior people in that building lean in, "why is the economic case for this entire industry resting on a 2001 survey?" is a very good one — and the freshest usable adjacent datapoint, the Uptime Institute's 2025 outage analysis, is worth having ready: power causes about 45% of impactful data-centre outages, 57% of recent major outages cost over $100k, and one in five cost over $1M.

(c) "The power quality market" is three different businesses wearing one label.

Different buyers, different channels, wildly different margins. Reasoning about "the PQ market" as one thing is the fastest way to be wrong. Also worth knowing: OMICRON, who you'd assume is a monitor vendor, isn't — they sell the test equipment that verifies everyone else's Class A claim. They're the referee, not a player, which is a nice position and worth understanding before you talk to their booth.

(d) Market sizing here is soft — and the spread is the real signal.

Published estimates for "power quality equipment" put the 2024 base anywhere from $20B to $42B — a 2× spread on the same named market, because the definitional boundary (does UPS count? surge protection? capacitors?) is doing all the work. Consensus growth clusters around 5.4–6.8% CAGR. The tightest and most credible slice is active harmonic filters at roughly $0.7–1.2B growing ~8%, where three independent firms agree within about ±25% on base and 0.4 points on growth rate — unusually good for this genre. Treat everything else as directional. And be suspicious of the standard TAM argument that sets a €150B cost-of-poor-PQ against a ~$30B equipment market: both numbers are weak, so their ratio is weaker still.

The category is consolidating right now, which is useful context for corridor conversations: Powerside acquired Electrotek — and with it PQView, the de-facto utility analytics platform — from GMC Instruments in March 2025, and private equity took a majority stake in A. Eberle in April 2026. Expect roll-ups pairing instruments with analytics.

Why the demand curve is bending right now — the five forces to have at your fingertips.

1. Data centres, in both directions. US data-centre consumption went from 58 TWh in 2014 to 176 TWh in 2023, with projections of 325–580 TWh by 2028. Inbound, AI clusters are the most sag-sensitive load class ever built at that scale. Outbound — and this is the new part — multi-hundred-megawatt loads with second-scale ramp variability have become a grid problem, and interconnection rules are starting to specify harmonic emission limits and validated load models for large loads. That turns PQ measurement from a nice-to-have into an interconnection deliverable with a named budget holder.
2. EV charging. Numerous, small, built by non-utility developers who often can't energise at requested capacity without a harmonic study and a filter. The segment most likely to want a cheap, standardised, self-service product — the exact opposite of the traditional utility sale.
3. Rooftop solar. Reverse power flow, overvoltage causing inverters to curtail themselves, unbalance from single-phase installs, resonance at the PCC. Fundamentally a monitoring-density and attribution problem, not a mitigation-hardware one.
4. Grid-forming requirements. ENTSO-E published its Phase II grid-forming technical report in November 2025 with a final report due Q1 2026, and NERC is running a parallel track. Every requirement in those specs needs high-bandwidth, time-synchronised instrumentation to prove at commissioning. A brand-new compliance-testing category, arriving on a regulatory clock. This is why TS 5 and Panel 3 are worth your Wednesday morning.
5. Supraharmonics — see A.5. A hardware replacement cycle waiting on standards convergence.

Three habits that will double what you retain.

1. Keep a running "words I nodded at" list. Every time you let a term pass without really knowing it, write it down. Look up five each evening. In a dense specialist field this beats trying to follow everything in real time — and the list itself tells you where the jargon walls are, which is useful product information on its own.
2. Ask the same question of many people. Pick two — I'd suggest "what's the most common power quality problem you see that people misdiagnose?" and "what would you measure if measuring were free?" — and ask thirty people. The distribution of answers is worth more than any single expert opinion, and it's the kind of thing you can only collect in a building like this.
3. Write it up Thursday night, not next week. Conference notes decay faster than almost any other kind. An hour on the train back is worth a day a fortnight later.