Two Weeks After Lift-Off

After a little delay our photovoltaic generator went online – we had been waiting for the delivery of this sophisticated addition to our office decoration:

Office Decoration

People on G+ had very cool suggestions, such as a rotating alien-fighting device throwing darts. Closest to the truth were: fuse box and fire alarm.

The box containing two knobs (actually this large box does not contain a lot):

Box with switches for PV DC cable

Two switches that are connected to that big red button downstairs, positioned next to the inverter for our PV panels:


We have two strings of modules, oriented perpendicular to each other; so irradiation on these is different. I add an overlay to a screenshot from Google Maps:

Plan of our house with PV modules.

Solar panels subject to different irradiance are connected in different strings – serial connections of modules; otherwise output power would suffer. The inverter has two inputs for two such strings and two MPP trackers that try to find the Maximum Power Point for each generator, by constantly probing each string’s current versus voltage curve.

Each string is connected to one of the little red knobs, which are part of yet another safety mechanism. The inverter converts DC current from the panels to standard 3-phase AC output voltage (230 V each phase). It has surge protection (another grey boy, but downstairs) and can shut off power at its DC and AC connectors – but then there is still a voltage drop across the DC cable from the roof to the inverter.

DC voltages supplied by our PV generators are about 400V, but generally they can be close to 1000V. This is a risk for firefighters connecting themselves to the circuit via a jet of water. You ‘cannot turn the panels off’ as long as there is sunlight! In order to make sure that the voltage drops to zero as close as possible to the panels, those switches are installed.

That ‘firefighters’ switch is semi-mandatory here. Lightning protection is not mandatory too, but we decided we should finally have one. Since safety standards and costs of such protection have grown exponentially in recent years, we can brag with a Faraday cage with tighter meshes and taller antenna-style tips than all our neighbours.

Alien EMP Protection

I am sure it protects us not only from lightning but also from alien attacks (see image below) and EMP guns – and the wiring goes well with the surface-mounted aluminium tube for the DC and AC cables for the PV generator.

Alien EMP Protection

The big red button is in the tech gadget closet on the left side of the driveway.

Firefighters will pull or push the red button in case of a fire. We decided for the pull option as you are less likely to pull than push something accidentally.

What we did not know before installation: The switch will also be activated automatically in case of a power outage – this means: about every 2 years for a few minutes. But when the big red button has been activated you need to switch power on again upstairs in the roof, too!

Normally, the switch box would be tucked away in an attic, above a dropped ceiling. We have no attic anymore – this is all office space, 3,5 high in the center. We could have squeezed the box into the insulation. But then after every power outage we would have needed to climb up there, remove roof tiles and switch on power again. So we spontaneously decided to have it installed on the ceiling, above the Chief Engineer’s desktop:

Office Decoration

Last Monday The Metering Guy from the utility finally installed a smart meter, capable of metering both consumption and feed-in to the grid. He had to disconnect from the grid to do so. We switched on the inverter in bright daylight – and there was no power! Panic – what happened? I fetched the laptop and the inverter’s manual, ready for troubleshooting – until The Chief Engineer walked by, carrying a ladder, and grinning mischievously:

Have you perhaps triggered the firefighters’ switch when disconnecting from the grid?

I had forgotten about the switch only about 15 minutes after I put up big signs for firemen! But at least we knew it worked!

After one more controlled test of a power outage we were finally online. This is what power generation looks like on a nearly perfect sunny day now (2015-05-11).

PV Power over Time, 2015-05-11

Since May 5 we have consumed 11kWh / day on average; about 55% of this have been provided directly by the solar panels. Daily energy generation was about 23kWh; we used 27% of the power generated.

25 Comments Add yours

  1. Cleon Teunissen says:

    Hi Elke,

    you mentioned that a so-called smart meter has been installed at your place.

    Given your background in computer security I assume you keep yourself informed, so most probably you are aware of the following article, but just in case I’m linking to it anyway:

    Practical Cryptanalysis of the Open Smart Grid Protocol
    Philipp Jovanovic and Samuel Neves

    Click to access 428.pdf

    Abstract: This paper analyses the cryptography used in the Open Smart Grid Protocol (OSGP). The authenticated encryption (AE) scheme deployed by OSGP is a non-standard composition of RC4 and a home-brewed MAC, the “OMA digest”.

    We present several practical key-recovery attacks against the OMA digest. The first and basic variant can achieve this with a mere 13 queries to an OMA digest oracle and negligible time complexity. A more sophisticated version breaks the OMA digest with only 4 queries and a time complexity of about 2^25 simple operations. A different approach only requires one arbitrary valid plaintext-tag pair, and recovers the key in an average of 144 queries, or one ciphertext-tag pair and 168 queries.

    Since the encryption key is derived from the key used by the OMA digest, our attacks break both confidentiality and authenticity of OSGP.

    1. elkement says:

      Thanks, Cleon! Yes, I have seen this article – and I also had wondered about the home-grown digest myself before when I read the specification of the OSGP two years ago ( Not using standardized, tested hash functions is typically not a good idea.

      You guessed right – I find this very interesting: I completed a post-graduate degree in renewable energy engineering in 2013, and I had written my master thesis about IT infrastructures and security architectures of smart metering solutions. Many protocols use(d) home-grown crypto or none at all. Risks might been mitigated though as many vendors have been working on solutions using real end-to-end encryption, using X.509 certificates – so all that insecure protocol traffic would be encrypted. But standards were still being developed at the time EU countries evaluated systems and pilot setups; so at the time I did that research it was unclear if a PKI-based solution would be feasible for all smart meters rolled out before the deadline.

      Our local distribution system operator has done pilots with a bunch of vendors, and now they have made their decisions. Meters will not be rolled out to all homes before ~2018 though. We just got this one because of the PV generator – and it is not really smart yet: The device as such is a smart meter, but lacking the infrastructure (data concentrators in substations and PLC communications) you cannot access the metered data as a consumer yet at the utility’s web portal. You can access them locally at the meter, if you invest in more interfaces and connectors – which we will most likely do, for detailed monitoring.

      1. Cleon Teunissen says:

        You write:
        “The device as such is a smart meter, but lacking the infrastructure […] you cannot access the metered data as a consumer yet at the utility’s web portal”
        ” many vendors have been working on solutions using real end-to-end encryption, using X.509 certificates”

        I gather that the smart meter at your place is in fact communicating with your power utility computer system.

        Janovavic and Neves wrote:
        “We note that our analysis has been performed solely against the OSGP specification and not against any deployed devices.”

        Have you been able to verify that the vendor chosen by your power utility has wrapped the vulnerable protocol traffic in end-to-end encryption?

        About a year ago I received a notice that the energy meter for my house would be replaced, as part of replacing the energy meters of a large block of houses. In the coming years all energy meters in the country will be replaced; apparently the block of houses I live in happened to be among the first wave. I specified that I didn’t want any of the smart features turned on, and The Metering Guy said he would do so. (Nagging doubt: I have no way of verifying that the device will in fact not communicate at all. Nothing I can do about that, I guess.)

        Am I overly suspicious here? Probably, but I don’t need the meter to be a smart meter; reporting my energy usage once every year is fine by me. I feel strongly about the importance of being alert about computer security.

        1. elkement says:

          I know for a fact that my smart meter is not yet communicating with the backend as the said infrastructures (PLC, data concentrators) are not in place at my site. I have discussed this with an expert involved in the rollout project – in addition to The Metering Guy who told me the same. Due to lack of that communication somebody from the utility will visit my place and read off data from the meter’s local storage once a year (using the IR interface) – not having to do this is one of the main returns on that big investment for the utilities; so I have no reason to believe two different people from the grid operator would lie to me. They were also very helpful with providing me specifications and manuals for the meter upfront – so that I could do research on the various logging options I would have.

          But my meter is just a temporary solution and the final solution that is decided upon just now will use meters by another vendor. Meters as the one I use now are only given to owners of PV generators on a case-by-case basis, and they will be replaced once more in a few year. The person in charge of the purchasing process said he could not yet tell me for sure which type of meter will be finally installed, so I don’t know which protocol it will use.

          Many of the meters on the market are also designed with extensions and ‘modules’ that might allow for a remote firmware upgrade (only the part ‘certified’ as per metering law must not be updated remotely) – so it might be possible that communication protocol will change even after the unit has been installed.

          No, you don’t need that meter even for detailed monitoring – you can simply install your own smart meter ‘behind’ the official one and use alll kinds of logging interfaces and protocols which are easier to access that trying to attach an IR sensor to the official one etc. I am interested in analyzing the peaks in power over time every few minutes (just as we do for PV energy production and all sensor readings from the heat pump system).

  2. Here’s a question for you: one of the reasons why PV is not used so much in my area is that we get relatively little direct sunlight. First, our high latitude (around 47.6 degrees) gives a worst max alt at about 20 degrees in the winter and the best is about 60 degrees in the summer. That’s not the big problem, though. No, it’s the cloud and fog. According to Environment Canada, during daylight hours we get sunshine about 35% of the time. I have no comparison figures for where you live but am guessing it’s significantly more. So here’s the question: how is the output of your system affected by clouds and fog? It would be interesting to see output daily energy graphs as a function of available sun, that is, all other things being equal, what would it look like on a day when you had full overcast, 30% overcast and so on.

    1. elkement says:

      We are at exactly the same latitude, so yes – the difference can only be clouds and precipitation!
      There are some free tools provided by the European Union that let you calculate daily and monthly PV energy forecasts and that take into account typical local weather. I will check if one of those has data for non European countries. Otherwise, I can at least check some weather tools providing global data.

      One thing I know from the top of my head is that direct radiation only amounts to about 50% of yearly PV energy gains here – the diffusive radiation on a cloudy day or when the panels are completely shadowed has a bigger impact that I had guessed.

      1. That leads to the small possibility that a certain light cloud cover might give a larger power output as it would better distribute the light energy to all three banks you have. I doubt it, though.

        1. elkement says:

          There is one interesting effect related to clouds – I have read about it before and now saw it my first logging data… and it is even more impressive than I figured: The output decreases with increasing ambient temperature (I-U-curve of the diode shifted), so you can expect the maximum output in ‘cold’ but sunny spring, not in summer. But when clouds cover the panels temporarily, the panels cool down – and when the clouds move along and sunlight hits them again… the output power is very high!

        2. elkement says:

          I found these numbers for St. Johns now (ugly link below) –
          Global radiation on a horizontal surface – which is the most common unit for an easy comparison of different regions:
          2.5 – 3.3 kWh/m2 per day –> 900 – 1200 kWh / m2 per year
          In Austria the equivalent is 900 – 1300 kWh/m2yr, and where I live it is closer to 1100 – 1200.

          But the whole island has the same color – can that be too coarse and not taking into account special conditions at the coast?


        3. elkement says:

          OK – found a book with data for St. Johns specifically now (Link below, table on p. 54):

          Yearly average:
          11,06 MJ/m^2d = 11,06 / 3,6 kWh/m^2d = 365 * 11,06 / 3,6 kWh/m^2d
          = 1121 kWh/m^2a

          About the same number as our yearly global radiation – and the ratio of diffuse / direct radiation is similar, too (6.17 versus 4.88, total 11.06 – checked data for our place again… more like 40% direct versus 60% diffuse, too!)

  3. That’s fascinating. I never figured the PV panels were configured to generate such a high voltage! I am guessing they’ve connected together a large group of cells in series to do that. I would guess that each cell is, what, about half a volt or so? I’m wondering why they don’t keep the output down a bit more for safety reasons. Also, it would not be too hard, would it, for the manufacturer to find an easy way to decouple some of the connections, on demand, again for safety’s sake, just for times when you want to cut output power.

    1. elkement says:

      Yes, the voltage per cell is about 0,6V (Si diode). The voltage is as high as possible to minimize resistive losses in the DC cables – it is the same argument as with transforming to high voltages in power supply lines: An array of cells is equivalent to a certain power (determined by ‘rate of photons per area’ at a certain irradiation), so power is given –>
      Losses are R_cable * I^2 = R_cable (I*U)^2/U^2 = R_cable * P^2/U^2.
      Losses can be an issue, depending on the position of the inverter. You would also try to find an optimum position where the sum of DC and AC losses is lowest. But since DC voltage is higher, you would go for shorter AC cables if possible.

      I think having additional switches between the modules would make the setup more complicated – and much more prone to failure of those. Now PV modules are equipped with short cables and plugs, so you can easily string them together.
      Actually, you can already have a small inverter per module – sold as ‘optimizers’ (as you have as many MPP trackers as modules)… but this makes the systems much more expensive. But the inverter(s) are/is also the component(s) that will fail first: The guaranteed lifetime is about 15 years for larger inverters, whereas modules will not fail in decades. I think I would not want to have so many predetermined breaking points.

      1. Of course! I should have thought of that :-) the fix is conductors with a larger surface area and, of course, steeply increased cost.

        1. elkement says:

          A large metal roof perhaps :-) !

  4. Michelle H says:

    Congratulations on yet another step away from grid dependency. Now in case of anarchy and regime collapses you will not have the standard excuses for closing shop, going out of business and hiding in the cellar. ;)

    1. elkement says:

      Thanks :-) The next step is to play with shifting loads to the sunny hours – e.g. turning on the washing machine or doing major updates or tests on a bunch of computers mainly when the sun is shining!
      It it odd but it makes me feel to be tuned in to ‘nature’ even more (in addition to gathering wild vegetables from the lawn.)

      1. Michelle H says:

        You are reverting back to the days before Edison and his light bulb. :) do you think it will be easy to do this with your night owl preferences?

        1. elkement says:

          I think there is a chance to reset my inner clock! From all the books and articles on sleep research I concluded that so-called night-owlishness or Delayed Sleep Phase Syndrome is due to an inner clock running a more then 24 hours cycle and / or low sensitivity to external cues resetting the clock. The solutions are 1) to run to clock as free as possible after you have changed your life-style accordingly – I think I have now exploited that option more than most people— or to 2) return to an ‘ancient farmer’s life-style’ with sunlight all the day and less light, especially UV light from screens, in the evening :-)

          1. Michelle H says:

            I think I also read that this type of sleep cycle adjusts more easily to sleep pattern changes than those of us who are wide awake when the sun rises but also feel regret signing up for committees and community service projects that require us to be out after 8pm. Our house here is fairly dark; even with new windows and a lot of tree trimming, we just don’t get a lot of light inside due to the roof overhang. I will say that I feel sleepy way too much of the day, especially in winter, and generally out of sorts from November until March. I think I am finally acknowledging that I can’t work from home anymore, at least, not from here. Reading much of the same sleep research literature, I either need to cut holes in the roof or go out most of the winter day to a sunnier location. I’m not the type of person who ignores these cues in light and seasons easily!

        2. elkement says:

          (again replying up here… the ‘notification issue’ :-))
          From experimenting a bit with habits I feel that for me it is definitely computers and especially doing something ‘interactive’ and ‘exciting’ with them in the evening. Cutting back on that helps me more than spending more time outside. TV should be bad due to the same type of radiation but perhaps is not as it uses a smaller angle of your whole field of view or because just watching passively makes you sleepy.
          Worst is maybe responding to clients’ e-mails – I recall that I was wide awake late in the evening when doing online troubleshooting, on days when I had to get up before 4:00 AM to catch a plane and then had worked a whole day (even in a dimly lit ‘nerds’ office’). And of course coffee: I finally lived up to one promise I made a long time ago on this blog – caffeine only in the morning.

          1. Michelle H says:

            I try not to check messages after a certain time, or I can’t rest for thinking about the response or the problem raised. Even after considering the mental distraction that work can be, I read somewhere that it is the blue light in our devices and computer screens that makes our brains think we’re still in day light, releasing a chemical that makes us more alert and ready for daytime activities. Too much of this sleep and chemical irregularity can lead to mental health problems later on(dementia), as these ‘awake processes’ don’t allow for necessary healing of the brain. (I have to find where that assertion came from; I wonder if it is even reliable, or if my memory is playing games with me…maybe Secret World of Sleep by Penelope Lewis?) There is a TED talk out there, too, which argues that people with mental illnesses have great difficulty regulating to the sleep patterns most of us have–that is, the standard 7 or 8 hours in a 24 hour period–and if those sleep patterns can be repaired many of the problems arising from brain dysfunctions can be reduced. It is all quite interesting. It leaves me to wonder what we humans might be like once we better understand the processes of optimizing our brains; would we all start acting smarter? But now I am far off topic of photovoltaic systems… ;) unless, we can find a connection between brain optimization and living within the limits of our solar harvests.

            1. elkement says:

              … maybe using a transparent solar cells cover for TVs / computers in the evening that harvest just the blue and UV light ;-) I have read there is software that changes the spectrum of colors of computer screens, reducing the intensity of blue and increasing red.
              Otherwise I agree with the author who conjectured that future generations might think about our tinkering with something so important as sleep (as per our ‘modern digital life-style’) in the same way we think about cigarettes or asbestos for insulation today.

  5. Brilliant, Elkement. Just to let you know that red buttons are seriously cool, whatever G+ says.

    1. elkement says:

      Thanks :-) Yes, we like them a lot :-)

Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.