Solar Zero challenge 3

Inspiration

It looked like the most exciting project from the three solar zero projects. It was exciting to try and build a model that big using real world data.

What it does

It models the power grid based off the paper provided. It can take in various parameters to specify the amount of different power sources it uses. It returns power used at very time step (which can be disabled), the total power from every source, and the total cost.

How we built it

We split the problem up into several smaller models which we divided between the team members on a kanban board. One of these was the reservoir model, which would provide a simple interface to the main model and keep all the complex logic about water arriving and leaving, spillways, and capacity hidden away. This allowed us to each work on different models, and one person create a main model which interfaced with all these but had much less complexity because it didn't have to worry about the nuances of different power sources.

Once this initial model was built, to get the lowest cost we created some functions to plot cost against different parameters, and used these graphs (along with a bit of trial and error) to minimise the function as best we could. We discovered batteries, in our model at least, were so cost inefficient that they are not worth buying, and using solar supplemented with wind works best.

Challenges we ran into

Understanding the model at the start
Re-learning python and different libraries
Dividing the project into chunks manageable by each person within their ability

Accomplishments that we're proud of

Getting our first graph with the signature "two hump" shape at about 2am, proving that our model was onto something.
Getting our model to reach, and not overflow, the demand of the grid