Brightfield GCP AI Interface

New product launch announced at Google Cloud Event

Nicholas Ord
3 min readJul 7, 2023
Adaire Fox-Martin, Google EMEA President — Keynote at Google Berlin Launch yesterday showing new Brightfield AI interface for already installed (offline) industrial energy meters

What does it do and why?

DC inverters translate renewable energy into 3 phases feeding the grid. They have to “chop up” the DC into AC so the grid can move it around.

This means fast swings, dips, swells and harmonics. If it were blood in the arteries; coronary heart disease. Energy cannot flow easily.

Heart attacks are often caused by blocked arteries and the number one health issue of transporting energy in the future is very similar.

Decentral two way renewable energy in power grids (made for one way coal/gas generation) is like forcing a heart patient to run uphill.

Diagnosing and predicting these issues in real time is essential to keeping energy flowing as wind, solar and cars scale.

Hardware and cloud data requirements for energy in English here and German here for handling future renewable energy transition

There are 450k transformers and industrial factories in Germany and about 20% are fitted with these offline meters. About 4 Mio in the EU.

Physically invasive installation is the main issue to digitisation. Some grids have as few as 50 qualified high voltage electricians per 20000 transformers.

Germany can only digitise about 500 transformers per year with standard equipment. 100k transformers will take 250 years at current rates. Less than 25% of the critical mass needed before coal and gas switch off in just 7 years.

With a few qualified engineers a city like Berlin or Munich can be online in real time with AI energy data within a year (4000 transformers and factories).

New sub-systems pull measurements at high speed from legacy analogue meters, process at the edge and convert it to AI data standards

In order for the cloud to get real time accuracy across the entire grid area the data standard IEC 61000–4–30 Class A defines exactly what kind of data and particularly how fast and how accurate it needs to be in clouds.

With launch planned for 2024, the new Brightfield product aims at this level of data quality. A low cost, fast upgrade to the existing installed base of energy meters for transformers, factories and renewable power plants everywhere.

Brightfield gets your offline energy data

(1) online fast in real time

(2) AI ready:

{clean X time axis in millisec (NTP)

50 parallel measurements Y axis per 150 millisec

X and Y synced to all assets in the field Z axis}

(3) Shipped from a German factory, secure tamper proof chip in every unit is scanned by installers; physically authenticated digital

(4) Your own Brightfield production cloud — terraformed to your company in a day. Your IP address — your data.

Early Testing with Beluk EMM5

for further reference see Google Cloud Blog

[chain code] 93e3 bebc c164 d766

Patents here