Earth Notes: Energy Game Changer: HaaS OSHUG Camp 2018 Talk

Why are clocks slowing down over western Europe like halving your heating bills at home?

Remember when at the start of the year a row in one corner of what used to be Yugoslavia caused clocks across Europe to slow down and eventually lose six minutes? Nothing technical was broken, and it's a reminder that people issues can't just be 'fixed' blindly with tech alone.

When a purely technical fix for energy efficiency is installed, for example a better boiler, savings tend to persist for many years, maybe for a decade if the tech lasts that long.

Solutions that may make as big a difference but rely on the people around it continuing to do something to assist, tend to have much shorter persistence. Maybe between one and four years.

We already have a smart radiator valve called "Radbot" that can knock 20–50% off your heating bills and pay for itself in a year. It requires very little input if any to do its job. But it won't work if people open all their windows in winter and expect magic to happen. Yes, some people do.

We just finished an Innovate UK project "Heating as a Service — Lite". We are more convinced than ever that while the technical and financial elements are probably easy to find solutions to, the social part, making things that work with real people for a long time, is intriguing!

Carbon, Radbot and Heating as as Service

My personal mission for the last ... cough ... years has been to cut carbon emissions and to reduce the impact of climate change. My current aim is to knock several percent off Europe's entire carbon footprint, and do it with a product that pays for itself in a year. And I would be very happy if it were to become nearly obsolete in twenty or thirty years' time!

Let me tell you about Radbot and the open software and hardware it is built on, OpenTRV!

Most of the housing stock that will be with us in 2050 is already standing, and most of what is standing now wouldn't meet current (inadequate) building regulations.

If you're an early adopter or a geek you may well enjoy upgrading your home with a system that you program in advance having worked out what your routine is, etc.

My experience with one such was illuminating. I have an MSc in theoretical computer science, and found what I had hard to understand and set up!

Radbot does not need programming, does not need the Internet, does not need a smartphone, and yet can save you enough money to pay for itself in a year, that's as much as 50% of your space-heating costs.

And I care that you save because you're also saving something like 200kg of CO2 per Radbot-equipped radiator per year. Did you know that there's around a billion radiators across Europe, half of them in homes?

Radbot is build on open software and hardware designs (see our GitHub repositories). We also publish a bunch of the data we use to calibrate the algorithms. I'll mention a bit later how we're the new Radbot available to the community.

People, Who Needs 'Em?

No one normal — in this room we're not! — wants to spend more time with their heating...

Novelty wears off, can we keep more energy-efficient behaviours?

Gains, Pains, Jobs

What about if you can make those energy-efficient behaviours something that also directly assists their main job?

Poor Folk Physics

Is heating a form of energy like lighting?

Do people react differently based on the scale of the units?

Can't Fix Anger and Apathy With Tech

Not even gamifying...

SPAM, 409, greed, venal and stupid walk with us; carrots and sticks and nudges?

Money and TRV2

Here to move that carbon needle, so need profitable growth company, investment, sales and manufacturing... And we have that!

But we also received lots of valuable pre-investment funding from the likes of Climate-KIC, Innovate UK, DECC/BEIS, and others. If you've got a good idea, especially in cleantech, there is lots of help and support out there. At least drink the free booze at the networking events, even if you're a bit STEM like me!

Test Often, Verify Reality Occasionally

Bosch's talk at FOSDEM 2018 on product testing with robotframework had it about right. When building hardware and IoT, model as as much as you can to make dev as much like software as possible with rapid comprehensive unit tests and CI (Continuous Integration). Then when you run real-world experiments you validate your algorithms and your test models. The models allow you to do orders of magnitudes more tests and tuning. And for those of us with heating products, we can develop even in the summer!

Why do we need to do all this testing for a boring radiator valve? Well, our code stack is something like 100kloc. That's bigger by far than the entire fixed-income library stack for Lehman Brothers when I joined them in the 90s!

For the OSH Community

We've made it possible for you to buy Radbots, use a dongle and flash it and you can play to your hearts content!

We're also looking to build an in-house controller with optional connection to the internet. Clearly will be Linux based. Looks like we'll have a job for an open sourcerer soon...

TRVs — rad through they are by definition — are not our entire universe. We expect over the years to develop other new easy carbon-saving tech for Jo Everyone alongside improving heating controls.