Add tutorial for stochastic generation simulation and desired averages

[harrison-lccc]

Task summary
Add a tutorial notebook under docs/tutorials/ that explains how the intermittent generation simulation samples historical weather/load-factor periods and how desired_averages reshapes distributions.

Why this task is needed
The simulation combines calendar-aware stochastic sampling with inverse-CDF resampling, but this workflow and its interpretation are not currently documented in an approachable, executable tutorial.

Proposed implementation
Create a notebook following the initial structure and conventions of docs/tutorials/wind_calibration_tutorial.ipynb. Use small, self-contained synthetic data where practical, introduce WeatherData and its RNG inputs, demonstrate historical-period stitching, explain histogram/CDF and exponential-tilt desired-average adjustment, and show the solar ignore_zeros behavior. Include visualizations and reproducibility notes without requiring private or downloaded datasets.

Acceptance criteria

• A new notebook exists under docs/tutorials/ and is readable from setup through conclusions.
• The notebook explains and demonstrates stochastic historical sampling and seeded reproducibility.
• The notebook explains desired-average adjustment, including inverse-CDF mapping and solar zero preservation.
• The notebook can be executed with the notebook dependency group and uses repository APIs where appropriate.

Additional context (Optional)
This is documentation/tutorial work based on the existing rencal.simulation and rencal.core.bucketed_data implementation.