#41 By: Graeme Stuart, August 7th, 2014 05:17
In IESD there are ~20 network switches split between two boxes. I'd say they are likely to average out at 100 - 150W each as they are unlikely to be particularly new. Queens has many rooms with wired network access provided similar to IESD.
We may be able to quantify it and explain a big chunk of the remaining baseload. @ibrokola is looking into it.
#42 By: Mark, October 15th, 2014 09:11
This is an interesting discussion thread and a good example of what the Smartspace project can do/show. I think the electricity baseload in all DMU buildings should be analysed to understand what contributes to the electricity usage. This could yield significant energy savings for DMU.
#43 By: Graeme Stuart, October 15th, 2014 13:07
We have made some progress with the investigation as detailed in the MSc dissertation of @ibrokola unfortunately I only have a hard copy.
The network switched turned out to account for more than 20% of the baseload if I remember correctly.
#44 By: Paul Fleming, October 20th, 2014 04:09
Does anyone have any further insight into the relatively high electricity baseload in Queens? I have noticed that the electricity consumption has been relatively high in Queens (all the time) since the start of this academic year. Does anyone have an explanation for this?
#45 By: Richard Snape, October 28th, 2014 07:49
As this thread shows - the baseload consumption is always high (40 kWh per half hour ~= 80 kW constant load) not just this academic year. Hopefully we can see a copy of @ibrokola dissertation soon - he has done a huge amount of work in order to identify the equipment in at least some of the baseload. The baseload does increase marginally in term time (as we might expect), but it's the peak load that increases most in term time - again expected as we'll see more use of discretionary loads such as computers in labs, lighting etc.
It seems that @ibrokola has done excellent work in allocating 20%+ of baseload to the networking equipment. Hopefully we can build on this to find the other 75-80% (~60 kW).
#46 By: Graeme Stuart, October 28th, 2014 09:54
These are the key findings from Ibraheems work. It shows the baseload consumption (~80kW) split by verified source. Nearly 23% of the baseload has been attributed to the network switches, 25% is the Creative Technology Studios and more than 33% was found to be lighting and PC's left on overnight. A good chunk of the unknown portion is likely to be from similar things left on overnight in areas of the building Ibraheem could not access at night.
Detail of the LAN switch calculation...
Here is a link to the full dissertation.
#47 By: Graeme Stuart, October 28th, 2014 11:25
In the five DMU smartspaces buildings the total baseload electricity cost is around £400,000
How much investment would be justified to reduce this by 10% and generate a £40,000 annual 'income'?
How much opportunity is there if we extended this to the whole campus?
How can we effectively put the buildings to sleep at night?
#48 By: Richard Snape, October 29th, 2014 07:47
Looks a really useful piece of work. The switches can even consume a bit more than that - the 740W figure is the Power of Ethernet (PoE) capability i.e. the power that can be supplied through the switch. The total power can be up to 910W per switch (if they are delivering max PoE). See page 25 of the spec here. Sad to go reading manuals, I know, but Ibrahim's work has inspired me.
Our phones consume between 2.4W and 4.9W each (page 4), so it would take some 145 phones per switch to draw maximum PoE. We might hazard a guess that the maximum power is not being drawn from all switches... It's good to have an order of magnitude idea of how much the switching and phone infrastructure uses, though.
It seems that automation of hibernation in network switches and computers, combined with intelligent switching of "accent" lighting (obviously retaining safety critical lighting) could make a reasonably large difference.
← previous page