Solar#6: 6 Months Post Install

Introduction

I kept putting off writing a 3 month review of our solar system - I want to write an honest appraisal and I had really wanted to be able to paint a 100% glowing picture of a fantastic investment...

...but, like most things, there is good and bad and the data is ambiguous. So, at last, I have bitten the bullet and written this 6 months review:

 

There are a few points that are not covered and I intend to cover in future posts:

• 6 months Return on Investment (ROI): as I write this, Octopus has resolved my smart meter issue and are just finalising my June through September bill. In this post I discuss energy (kWh) and power (kW), but in the end it all boils down to money (£)! A future post.
• Scheduling: GivEnergy are promising a solution to the firmware issue that, in some cases, means that scheduled discharging is not optimal - I do have a summary of this issue below and I will discuss the scheduling topic in more detail in a future post, once, I hope, the issue is resolved.
• Emergency Power Backup: I have not yet paid the extra for this facility to be installed.

 

I don't want this introduction to sound like I am disappointed with the system, so a few, more positive, points:

• It works - the system is robust and the 6 month performance looks to be on track to beat the 12 months figures in the Lovatts proposal
• GivEnergy software is cool and the overall hardware and installation looks neat, well designed and with solid construction.
• The GivEnergy software does allow the system to be scheduled for the Octopus Flux tariff and, so far, I believe it has been broadly optimal over the course of the spring and summer.

 

Lovatts Installation

All functioning well - solar system and roof.

Installation - My View: It Looks Good

Since the installation I have had not really needed any help from Lovatts and I would certainly recommend them to others. Their one shortfall is their limited knowledge of scheduling the GivEnergy system and they have tried to help here, but are limited by the support & documentation provided by GivEnergy.

 

GivEnergy Hardware

Not much to say here - it still looks good and has not skipped a beat.

 

GivEnergy Software - Phone App

This has great monitoring capabilities; this does help in understanding where energy is consumed and how to make savings - see section below "Consumption Visibility Gives Savings".

 

It took me a while to work out that the scheduling features of the phone app are limited; really a sub-set of the cloud app. Since working this out I have not used the phone app for scheduling. To add to the confusion:

 

• My installer had limited knowledge of how to schedule the system
• The GivEnergy helpdesk, while staffed by very nice and helpful people, do not have knowledge of the scheduling features of the system and they were not aware that the scheduling interface of the iOS app is different to the Android app version described in their documentation.
• The phone app documentation is limited: the more obvious features are fully covered, but not scheduling.

 

My view now is that the phone app (iOS and Android) are great as a monitoring tool, but it is best to do the scheduling in the cloud app. This is actually fine if, as I do, you want a set and forget system. It is easy to do a quick check on how the system is running on the phone app, but setting the schedule is a more considered activity, done (I had hoped) just once.

 

GivEnergy Software - Cloud App:

This has similar monitoring features as the phone app and much better scheduling capabilities. The cloud app also has some interesting reporting and data download capabilities that I have only made limited use of. There is no documentation for the cloud app (or none that I could find) and the helpdesk know very little about the scheduling features.

 

Scheduling with the Cloud App: Once I have the GivEnergy firmware bug fix I will report on this in more detail, in a future post. Right now my schedule makes use of the Octopus Flux cheap rate and peak rate periods:

• Timed charge: from 2am to 5am to 100% (Octopus Flux cheap rate)
• Time discharge 2 x 45  mins during the peak rate of 4pm to 7pm. The 2nd discharge finishes at 7pm. Both are set with a 4% minimum battery charge. My aim is to avoid consumption from the grid at peak rate and then to maximise discharge to the grid, without manual intervention.
• Outside of these times I use the ECO mode that works as you would want - ie prioritising home consumption of solar energy, then charging the battery and finally if home consumption is satisfied and, the battery is at 100% charge, it exports to the grid. If the available solar power is not enough for home consumption then the battery is used and, if this is exhausted, then the grid is used. If peak demand in the house exceeds the inverter/battery peak limit (nominally 5kW)  then the grid is used.

So far the battery has not been fully discharged at 7pm, on most evenings, ie we have only had very limited import from the grid at peak times and almost always because we have exceeded the peak capability of the inverter (nominally 5kW). This can happen, eg if the oven, grill and kettle are on at the same time, but actually accounts for very little energy (eg from my Octopus bill for 1st July to 28th Sept peak rate consumption is only 2.0kWh).

 

A bug in the GivEnergy firmware means that if we do not hit the 4pm peak rate window with the battery at 100% charge, and I do not adjust this time based discharge schedule, then we could end up with more peak rate import before 7pm. I think that this is likely to happen now that we are past the Autumn equinox.

 

Performance Numbers:

• Solar energy generation
• The Lovatt's proposals, based on the MCS calculation, states that for a 12 solar panel system we would produce 3,610 kWh each year. I was told to pro-rata this number for the 14 panel system, ie 4,212 kWh each year. After losses of 181 kWh/year (pro rated from the proposal), this gives 4,031 kWh
• For the 6 months from 1st April to the end of Sept the actual generation was 3,362 kWh. From what I can see the Solar generation energy figures given by the GivEnergy app are after losses, so 83% of the full year 4,031kWh in the proposal. It looks like we should exceed the proposal, by a decent margin, unless the winter is really, really gloomy!

 

• Power: The peak power of each solar panel is specified as 400 Watts in the proposal, or 5.6kW for an array of 14.  On 28th June the GivEnergy app showed a peak power generation of 7.257kW. I am a little dubious of this figure as on days with very variable levels of sunlight the peaks were always higher than on days with continuous sunshine. For example, the day with highest energy (36kWh) had a peak power of 5.38kW. I am not sure if this is an artifact of the measurement system of if the system is more efficient when there are only short bursts of high sunlight levels. In any case, given the non-optimal direction of the house, pitch of the roof and location on the globe, even the 5.38kW seems pretty good.

 

• System and Battery losses: I have struggled to make sense of losses - this is what I have worked out from the energy figures provided by the cloud app:
• Grid to home - there are no losses here (& you would not expect any, it does not go via the GivEnergy system).
• Solar: clearly there will be losses in the inverter, but the figures quoted are after losses, so it is not possible to work out (from the energy data) what these losses are.
• Battery in and out: This averaged 11% of the energy input or 1.28kWh per day over the 6 month period. During this time the battery was fully charged on most days, using cheap rate electricity and around 60% was discharged to the grid at peak rate with the remainder discharged to the house. On gloomier/shorter days there were one or more mini-discharge charge cycles. On longer sunnier days then the battery pretty much stayed charged and the house ran off solar/battery from 5am. While there is variability in the daily losses, there was no obvious relationship with the charge/discharge pattern on that day.

 

Solar production Seasonal Observations

Considering the 6 months from 1st April to 20th Sept (the first 6 months period where I have complete monthly data):

• Total production is 3,360 kWh, or on average 20.2kWh per day.
• The month with the highest production was June at 690kWh or, an average of 23kWh
• The month with the lowest production was September at 390kWh or, an average of 13kWh
• The June daily solar production range was 8.82kWh to 36.24kWh
• The September daily solar production range was 1.48kWh to 24.51kWh

It will be interesting to see how solar production fares in the winter months (as I make the final edit of this post, in mid-October, we have just had a miserable truly Lancashire day of  rain and dark grey clouds with only 0.76kWh of solar production for the whole day.

 

The thing that I have been surprised by is the enormous daily variability:

• Over 4:1 best day Vs worst day in the summer
• Over 16:1 around the autumn and spring equinox

Given this variability, then if a tariff requires scheduling to achieve maximum ROI (ie anything other than a flat rate tariff) then I think that the options are:

• Daily manual adjustments depending on expected weather and energy use profile
• A fixed schedule that is robust to a wide range of weather types and usage patterns
• A schedule that automatically adjusts depending on the weather forecast and expected usage patterns.

This topic is something for a future post - suffice it to say that my objective, at present, is the fixed robust schedule, & I think that, other than the bug mentioned above, the GivEnergy system allows this when combined with a tariff like Octopus Flux.

 

Accuracy: GivEnergy Vs Smart Meter Energy Readings

I took grid import and export readings roughly every 2 weeks from the smart meter and the GivEnergy software.  The readings for both import and export were in the range of 0.8% to 2.5% difference, with the smart meter giving slightly higher readings for both import and export. I was pleasantly surprised by how close these figures are and how closely the graphs match over complete days. The biggest deviations that I saw were in a period in September (9th to 20th Sept) where I was running the system to minimise grid import; on days where the import was <1kWh then the error rose above 2.5% - for example

	GivEnergy Cloud app	Octopus App	error (kWh)	error (%age)
11th Sept	0.17	0.30	0.13	43%
18th Sept	0	0.04	0.04	100%
19th Sept	0.29	0.33	0.04	12.1%

I am not sure that there is too much to draw from this other than there is a small error between the two meters that consists of:

• A small gradient error of <2.5%
• A small offset, noticeable at <1kWh/day
  
• A small amount of noise, noticeable at <1kWh/day

The smart electricity meters that I have had have both been labelled as 'class B' and apparently this means within 1% accuracy, according to IEC62053-21/-22 (but I must admit that I have not read this standard).

 

GivEnergy Cloud App May Export	Octopus Phone App May Export

 

GivEnergy Cloud App September Export	Octopus Phone App September Export

 

For both import and export the smart meter has slightly higher readings than the GivEnergy system, so no particular bias in Octopus's favour or mine (Octopus win a little on import, I win a little on export).

 

My view is that, even with these larger %age errors on low readings, the error band is pretty good and not significant in terms of billing.

 

Consumption Visibility Gives Savings

The GivEnergy phone app gives great visibility as to the instantaneous production and consumption in our home. My view is that, in general, I want to live my life without constraints set by energy consumption, but, having said that, if some easy adjustments mean that we save money then I am all for it.

 

Home Screen: instantaneous production and consumption	Power Graph: 3.3kW peak production - not bad for October	Battery Charge Graph

So how has this helped:

• The first quick win was that we saw that the dishwasher uses a little over 2kWh for each run - so we save around 20p each night or £70/year by remembering to run this overnight at cheap rate (it has a built in timer) - and is usually convenient to have clean dishes each morning (of course, this raises another question on whether to run a half full dishwasher over night! - we usually do).
• Anything with a heater is a high consumer - tumble drier, washing machine, iron, oven...
• Peak rate starts at 4pm and we charge the battery overnight at cheap rate. Export before 4pm gives us around 15p per kWh, export after 4pm around 25p per kWh. So if we get to 4pm with a full battery then we can export most of the contents at 25p/kWh. If we get to 4pm with a less than full battery then we might lose some export at 25p/kWh. So, do the washing on a sunny day, in the morning, then it is likely that you still get to 4pm with a full battery, and we have exchanged laundry for exporting at 15p/kWh, but do the washing on a gloomy day and the battery is not fully charged at 4pm and we have exchanged doing the laundry for exporting at 25p/kWh...

...Keep the laundry for sunny days - perhaps saves of 1 to 4 kWh (10p to 40p), or a £few per year.

• Before going on holiday it is possible to drop the background run rate consumption from 200W to 300W, by around 100W, eg by turning off various low consumption devices like the TV, printer, etc at the wall socket. 100W is 2.4kWh per day, so over a week's holiday that is 16.8kWh or around £2.50.

These are just some specific examples of how the visibility to consumption provided by the GivEnergy app is useful; being conscious of instantaneous consumption can be useful but, in my view, is unhealthy as an all consuming lifestyle choice! Spotting the bigger savings is perhaps worth around £100 per year for us.

 

Mature Technology? Where is this Technology on the Market Adoption Curve?

There is a marketing concept for new technology of the adoption curve

The Technology Market Adoption Curve

The concept of the 'chasm' was introduce by Geoffrey Moore in his book "Crossing the Chasm". My summary of the concept is that many technology innovations look like they have made a promising start, but actually it is the early adopters and visionaries who are buying them. Many technology products fail to convince the early majority to buy in - these people are looking for proof from other users whom they know and trust (ie not fancy marketing). There are plenty of people who disagree with this concept for all technology and for specific technologies. For example, for solar energy, does government incentive (eg no VAT) impact on the adoption curve?

 

My view is that the adoption curve is a useful concept, but is not precise and it is often hard to slot each individual buyer into a neat category.

 

Having said that, it is clear that for a technology product to succeed it must be easy to use and be seen be obviously useful if mass adoption is going to happen

 

My experience, after installation and 6 months with this system, is that solar panels, without a battery, are ready for the majority market and, as long as the numbers stack up, and you have a reputable provider, they are probably a reasonable investment.

 

However, add in a battery and the system is, in my view, very firmly in the innovator/early adopter phase. For a battery to reach its potential return on investment (ROI) it requires:

• Scheduling to get the full return on investment
• Flexible electricity tariffs and this needs a working smart meter (see my posts Smart Meter #1: My Journey and Smart Meter #3: My Journey Part #2).

Both scheduling and getting my smart meter to work has taken more time and effort than, I would have thought, the majority of people would want to spend.

 

My take is that I have selected (perhaps more correctly 'stumbled upon') two of the leading suppliers in the field (Octopus and GivEnergy), but it could be that other suppliers have these issues resolved - it would be interesting to get feedback from others on this point.

 

I do think that both of these issues are in the hands of the industry and, perhaps solutions are close at hand, if the players have a desire to resolve the issues. However, until the issues are resolved, my view is that domestic battery systems should only be purchased where the user is prepared for the effort required as an early adopter.

 

Summary:

The system is operating reliably and the software is good, with the exception of the scheduling points above (& more detail in a future post) and the issues with the smart meter (now resolved).

 

I am also coming to the conclusion  that the domestic solar+battery energy industry operates as a 'cottage industry'. It works well for enthusiasts, prepared to invest time and effort and is not really ready for mainstream users...

 

...but it is probably the case that the domestic solar (with no battery) makes life much easier, is a more mature market and is ready for the mainstream.