Goes without saying that the market has done a lot of shifting between the two years. When you separate out that anomalous March data the installation statistics indicate that 37% more systems sized up to 5 kW, 28% more systems from 5 to 10 kW and a full 50% more in the 10 to 15 kW size range are getting installed in 2013 compared to 2012. I'm not trying to lie with statistics here - the shift to smaller systems has certainly resulted in a significant reduction in MWs installed. That said there are still a lot of systems getting installed every month. Those systems represent jobs. It's nice to know the jobs haven't really gone away - They've just moved to working on smaller systems which don't get installed as quickly as the multi-MW solar farms. I really don't consider losing the solar farms a tragedy.
Interestingly, 70% of all the systems going in every month are under 15 kW. It stands to reason these systems are typically going on homes and businesses. I think the reason why this market segment has been so resilient is because these people are using photoelectric systems to protect them from having to buy electricity - electricity which has been going up in price by an average of over 5% a year for the last 10 years running. The wonderful thing for this market segment is that the incentives driving the growth (high electricity prices) are actually going up. The FiT rate can literally go to zero and the sub 15 kW market segment would still survive. Which leads me back to one of my favorite topics.
Some silly people seem to think that Germany is installing too much solar. These jokers think Germany should focus on wind. They use this argument that Germany will eventually have too much solar during the day and the grid won't be able to handle it. Here's a chart that's been used to explain the concept.
This projection tells you something about what current and future solar production will look like but it does a terrible job of projecting what load will look like. Here's a delightfully crude version of what I think things will look like.
Peak load will get peakier and baseload will shift down. Why will this happen? Simple... If it's 3 times cheaper to use photoelectricity compared to to buying electricity from the grid you can be damned sure that people are going to self-consume as much of their production as possible. When you've got 5 to 10% of homes and businesses all acting in unison to coordinate appliance runtimes with photoelectric production you're going to have peakier load. Interestingly, the grid operators won't see the peakier load - they'll actually be seeing flatter load at the network level. You have to imagine that whole neighborhoods might be producing 100% of their own electricity from solar during parts of the day. The feeder at the substation would see zero load under this condition. Lots of feeders communicating with the system control center would be showing these reduced load conditions so the network operators would see flatter load. It's fascinating when you think about it. I'm sure the network operators will figure out a way of determine what the shadow load is because this information would be important for maintaining system reliability.
I'll close with the basic reasoning behind the reduced load at night. If your water heater and fridge have already made all the hot and cold they need during the day you'll see a clear reduction in load at night - that's why I shifted the whole load profile down and shrunk the baseload requirements. This is important in regards to maintaining capacity requirements in Germany. The silly jokers who think they understand Germany's Energiewende seem to think that the grid of the future will need to carry enough firm capacity to meet the peaks in demand that we currently see in the fall and winter. These people aren't seeing that all the smart machines that can be used to shift load into the day for self-consumption purposes during the summer will be available to shift load into the night during the winter. That will have the effect of lowering peak load and reducing capacity requirements. A water heater doesn't care when it makes hot water so long as it has a big enough storage tank to ride out the expected demand swings. The reasoning in the SMA article I recently linked to is telling. These guys don't see just a few appliances getting smart - they're all in. They see a future where absolutely every machine is interfaced. Ya know what... they're right. This massive amount of interfacing will create a massive amount of virtual capacity. You wouldn't need to carry firm capacity for the peak in November. You'd reduce the peak in November using load management and then carry that capacity - that reduced capacity mind you. How much would it cost to give an interface to every machine with a plug? Not a whole lot it turns out... The video below suggest that each appliance would need a 10 cent digital transponder and power outlets would need to be upgraded with the addition of an "inexpensive wireless reader".
There are a couple of other wonderful things that come along with the deal. In this future where every machine is interfaced you'd save thousands of lives, prevent tens of thousands of injuries and prevent billions in damages done by house-fires. The video talks about it. Pretty amazing.