Take the power* back! (*monitoring and control at outdoor events)
I only recently saw that Firefly power have wisely created a Timed Distribution Unit for hire: (link) and it reminded me of some thoughts I had about a similar system a couple of years ago. So here it is!
AKA "How to design a power monitoring and control system that could be deployed in a festival environment, and what some of the benefits could be."
Managing power consumption and reducing energy emissions at a festival is not quite as simple as doing the same thing at home. This is mainly because it is impossible / far too costly to implement a proper power grid. At home you can tap into as much power as your circuit wants and fuses can take, you can do it at the drop of a hat and pay as you go for pretty much the exact amount (kilowatt hours/kWh) of power used. Plus if you are generating power, you can feed it into the grid at get paid for it (though the exact tariff varies depending on the mode of production, as well as the government policy at the time). Finally, diesel generators (even at their most efficient) produce approximately 50% more GHG emissions per kWh than the fuel mix from the UK mains grid. This means that efficient practices have an even greater impact (per kWh) in terms of reducing emissions than they would in other work or home environments.
Whether you are running a stage or a burger van, at a festival you typically request the maximum potential current you need in amps. At single phases, this is usually 13, 15, 16 or 32 amp with 63, 125, 400 and 600 amp more commonly being three-phase supplies. (No, I’m not going to explain the difference. Something about a shared ground connection. I’ve wired up light switches in my house and that is the extent of my abilities!)You usually have access to this power 24/7 whether you use it or not and in some ways therefore have limited incentive to both reduce and manage your power consumption. You might be able to add some renewable generation to your festival stall or trailer, but you won’t see any financial benefit if your peak power consumption remains the same, and there’s certainly no mechanism for selling or storing excess power either. (Maybe charge up peoples phones here and there if you like) Power to traders may run for 4-5 days maximum over the course of the main event, whereas other production areas may require power for around a month for the full build/get-out process.
The ESCObox project is a De Montfort University project which works to provide smart-grids for the developing world. Using relatively cheap electronics such as remote-switching power sockets that can be controlled (via infrared in this case) by a Raspberry Pi or similar microcontroller, it is possible to control individual sockets (in this case 13 amps) being fed from small scale generators. This means power hungry devices or ‘anchor loads’ need to be managed more directly to prevent grid failures. If your solar panels can only power one washing machine at a time, you need a way to ensure only one house gets to use their washing machine at a time (for example).
You could trust that individuals stick to a specific timetable, but adding a central controller means that this system becomes more foolproof. In general, mobile phones or small computers, or LED lighting, draw so little power as to not really be worth directly managing. The anchor loads (lets say, anything from 250w to 3kw per device) among festival traders are likely to be: water heaters/kettles/tea urns, blenders/coffee grinders, non-LED lighting circuits (halogens, some CFLs), small scale soundsystems/PA and cold devices (fridges, freezers).
In theory…
Imagine you are a festival power provider and due to the way the site is planned, you have to hook up six traders to one generator. Each trader (for simplicities sake) has a 10 litre tea urn. Each uses a 3kw element that runs at full power for around 25 minutes to bring the water up to boil, then goes into ‘standby’ mode using about 10% (300w). We’ve got a peak of 18kw (3kw x 6 traders) to deal with, if they all decide to turn on their tea urns at the same time. In the conventional approach, you would supply generator capable of 18kw peak, even if the peak capacity may only be required a few times a day, if ever. If we can control the times at which these anchor loads are available, we can reduce the peak required across these six traders and give them a smaller generator. Is this feasible without interfering too much with the general operations of the traders? Maybe!...?
10 litres is about enough for 40 cups of tea (250ml each). According to the specs, the peak heating time is ‘only’ 25 minutes (let’s say 30 to give a margin of error). We could allocate each trader a 30 minute window during which time they can refill and heat the tea urn. With six traders, this makes a 3 hour cycle. In this case, each trader could serve 1 cup (on average) every 4.5 minutes before running out of hot water – their 30 minute window comes around again and they fill up and heat another 10 litres of water. How many cups of tea does a trader serve per minute on average? Up for debate but 40 cups every 3 hours sounds… okay?
It’s worth noting in this example we’ve actually made it pretty hard for ourselves by specifying quite a large number of traders to manage, and only specifying 10 litre capacities. Larger tea urns have similar ‘peak power’ elements and although they do take longer to heat up the full volume, they use a similar amount to keep hot once boiled; obviously this extends the ‘window’ we have to work with. With a 20 litre capacity in this example, that becomes 2.25 cups per minute – one every 30 seconds or so. We would likely have more of a margin on the generator and timeslots – for example being able to have two traders heating up at the same time would mean the timeslots could overlap, maybe each trader would get 1 hour instead of 30 minutes, and their slot would rotate back around faster too.
And so on and so on. Lots of variables. Anyway, it’s the principle of demand shaping that is important.
The net result in this (extremely weird) scenario would be that instead of a 18kw generator, you could theoretically run a 3kw generator to achieve the same result. (For reference, that’s a purchase cost difference of ￡5000+ to maybe ￡500+). Even if you ‘only’ managed to halve the size of the generator to leave a substantial margin of error, it’s still going to be a good saving. From the generators point of view, it is far easier to predict when loads will come and go and much easier to match appropriate capacities to appropriate loads. Generators also run more efficiently as close to their rated value as possible (something like 80% of the maximum) so aiming at this target gets a bit easier too. From a fuel point of view you have a 3kw generator running at a high efficiency for a lot of the time, rather than an 18kw generator running at very low efficiency most of the time.
In this scenario we have created a more flexible market for temporary power. Maybe you could really work on insulating your tea urn? Or store excess hot water in thermos flasks/vacuum pump pots? Changing your kitchen layout to keep hot items (cookers) away from cold items (freezers) Everyone wants to use power at the same peak time? These slots become more expensive or could even be sold/transferred between traders if it turned out your demand was higher or lower than initially expected. If you can reduce your base load with renewables, or charge a battery to help meet these loads, you are incentivised to reduce the number of ‘windows’ you would need to buy and potentially, in future, are even able to sell or store excess capacity.
If traders only pay for power consumed, the power suppliers are the ones stuck with the bill for wasted energy. Given traders will cover many festivals and events in a year, it will clearly take time to change industry practices, but in future, rather than sending a request for a maximum current (amps) both parties could be using more detailed energy profiles.
Hardware
What are the ‘brains’ of this system? While it seems quite an easy task (on paper) to provide automated socket switching, there are many options available. You could just have someone manually turning things on and but that is not much fun! (Well, for some of us it might be)
We only have to look at behavioural studies to realise humans aren’t great at using basic systems like thermostats, timers and manual socket switches. Microcomputers like Arduino and Raspberry Pi are sufficiently powerful enough to run a basic ‘timetable’ loop that would send a signal to the switch controlling each individual socket. Advancing this on the software side further would even allow other intelligent features, like predicting loads (based on the weather and renewables?), sending warnings and powering up reserve capacity, monitoring truly ‘smart’ billing.
Control signals could be sent by Infrared (like a TV remote), WiFi or wired Ethernet. The IR used in Energenie is rated for about 30 meters and requires a clear line of site. Stock wifi antennas are a similar range and consume more power for both receiver and transmitter. Wired Ethernet can range up to 100 meters (in theory) and injectors can boost the signal further. Premade remote power sockets are available from companies like Olson though it is difficult to predict exactly how many and what type of sockets are required. Robot Electronics supplies a range of 16amp relay switch boards, ranging from 2 inputs to 20. These can be controlled by Ethernet and sockets could be added in whatever configuration required from 13-16 amps. Control Anything also have a range of automation products, including slightly cheaper timed relays – eg: set it up once, then just leave it running, though you can’t control it remotely of course. (And someone might tamper with it I suppose!)
The Gridtogo hybrid units from Off Grid Energy are rated from 5-30kva and the battery storage can range from 10 to 300kwh. (Those from Firefly are similar)These can be paired with conventional diesel/bio-diesel generators and renewable sources. It is available in different configurations, though the one pictured on the website seems to have at least 1x 32A out and 4x 16A outs. For reference the Powerpod has 750w of panels (6x 125w) and three batteries in series for 18.2kwh of storage (1560AH at 12v). The wind turbine (Rutland 913) potentially adds another 250w peak (90w rated) and pedal power can be added too.
Overall, depending on the exact deployment, it seems feasible to run a small ‘hub’ of traders with a comparatively small generator while continually ‘topping up’ from solar, wind or pedal power sources - all with some amount of battery storage capacity. Demand shaping requires a good deal of on the ground work from ‘middle men’ (like Face Your Elephant volunteers) in discussing the various controls and ramifications of this kind of approach – and suggesting ways the users could improve and save money.
Interesting stuff, if I do say so myself.