The homestead2 solarPV installation supplies a daily Power vs Time graph.
This graph is loaded as an image into Webplotdigitizer on a Chromebook. A set of representative points are selected on the graph such that linear lines between those points roughly outline the actual graph.
The linear interpolation between these data points is used to pickup kW at the hour markers to match the Hydro dataset.
Once a dataset of kW at hour markers is generated the kWh values are calculated using the following formula:
kWh(n to n+1)=(kW(n) + kW(n+1))/2 * 1hour
The result is a dataset of kWh for each hour of the day. This matches the dataset produced by electric utility. The PV graph also comes with daily total kWh. Our model bar chart can also produce a daily kWh total. The actual daily total is compared to the model daily total and a correction factor is calculated as:
Correction factor = Actual daily total / Model daily total
This correction factor is then applied to each model bar to bring the model to same daily total and the PV instrumentation supplies.
Model Assumptions
These datasets are incorporated into a Google Spreadsheet.
We are making the following assumptions concerning the storage algorithm in the model:
1)The PV output is first used to directly offset the home electricity load
2)Any surplus PV is allocated to storage without any losses
3)Any deficit in PV is drawn from storage without any losses
These assumptions (particularly 2 and 3) are justified because any losses in the storage can be readily accommodated by simply adding more PV capacity.
Model Algorithms
The homestead1 electricity and homestead2 PV datasets are captured daily. They are added to a pair of tables with 24 columns representing each hourly interval throughout that given day.
The hourly accounting for storage simply follows the formula:
To storage(kWh) = PV(kWh) - home(kWh) if PV>home
Or
From storage(kWh) = home(kWh)-PV(kWh) if PV<home
An hourly running total of the amount in storage is then calculated from the start date of 2023Nov1.
Since this experiment is being run throughout the winter months where there is high GSHP load the model includes an input variable which represents the kWh in storage on Nov 1 (presumably added there from summer surplus). We have confidence that this surplus is possible because previous years of homestead2 PV performance data has shown that the annual output is
1MWh/yr/kW on a 16kW solar PV array
homestead1 consumes
12MWh/yr
This input variable is simply adjusted to prevent the running storage total from going negative at any hour.
This kWh in storage on Nov 1 thus represents the answer to our experimental question; ie. how much seasonal storage would be required.