How SunSolve P90 works

Basic architecture

SunSolve P90 has two main components:

1. a Jupyter Notebook that runs on your local computer, and
2. the SunSolve engine, which runs in the cloud.

The Notebook acts as the user interface. It provides a structured workflow with pre-built Python scripts that you can customise to meet your analysis requirements. You can modify the input parameters and send them to the SunSolve engine via an API. Once the engine completes the uncertainty calculations, the Notebook scripts receive and display the results.

Users familiar with Python and Jupyter Notebooks can set up the environment on their local computer. We provide guidance here on installing the required development environment (e.g., a standard Jupyter Notebook, VS Code, or Jupyter Lab) to get started.

For beginners, we recommend using Google Colab to run the Notebook. As implied by the image below, Google Colab is a ‘wrapper’ around the Notebook, removing the need to configure Python development environments or supporting packages. All you need is a Google account and Google Drive, and it runs in your browser.

Our getting started section provides a step-by-step guide to running SunSolve P90 in Google Colab.

Contact us if you’d like to access the engine directly through the API.

If the tool proves popular, we’ll build a browser-based user interface to simplify the user experience.

Notebook workflow

We provide a default Jupyter Notebook that has been configured to conduct the analysis in 11 scripts:

• Scripts 1 & 2 log in and connect the Notebook to the SunSolve engine. They need only be executed at the start of a session.

• Script 3 loads the weather and base inputs for the simulation.

• Scripts 4 & 5 visualise the weather and probability distributions.

• Script 6 loads the uncertainty distribution.

• Script 7 runs the simulation.

• Scripts 8—10 display the results in tables, histograms and other graphs.

• Script 11 exports the results to Excel.

At the start of a session, it’s best to run all scripts sequentially. For subsequent analyses, it is not necessary to run all 11 scripts. For example, you might only want to modify an input with Script 3, rerun the analysis with Script 7, and export the results with Script 11. You can collapse the other scripts to simplify the navigation.

The flexibility of Jupyter Notebooks allows experienced Python users to modify and add scripts. They might create sweeps, display alternative plots, or integrate SunSolve P90 into their own workflows.

Follow our step-by-step guide for more detail.