PV-ABM-LCA

PV-ABM-LCA is the pilot integration of an ABM and an LCA based on two preexistent models: the LCA on III-V/Silicon photovoltaic (PV) panels from Blanco et al. (2024) and the ABM on end-of-life crystalline silicon (C-Si) PV panels from Walzberg et al. (2021).

How to Run the Model

1. Create a new environment: Use the file abm_mesa.yml to create a new environment with the command: sh conda env create -f abm_mesa.yml
2. Activate the environment: sh conda activate abm_mesa
3. Modify the metamodel trigger: Open ABMCEPV_Model.py and modify the trigger on line 10 to specify whether to use the climate change or resource depletion metamodel.
4. Configure batch run parameters: The defaults reproduce the results plotted in batch_visualize.ipynb. However, to overcome memory issues multiple runs are used and they can be reproduced by changing the triggers.
   • Open batch_run.py
   • Set the run type (fixed seed and multiple threshold or 100 seeds and 1 threshold)
   • Set step and sample_size for the fixed seed run:
     • step: Determines the %increment between consecutive values in the thresholds list.
     • sample_size: Specifies the number of values in the generated thresholds list.
   • This choice will generate two lists for the concern/indifference thresholds with length=sample_size and % increment=step. Max threshold of concern = max impact and min threshold of indifference = min impact.
   • The max impact is found by running the ABMCEPV_MultipleRun script without thresholds (thresholdconcern=high number, thresholdindifference=0)
   • The min impact is found by running the script batch_run with concern threshold but without indifference threshold (thresholdconcern=[maximpact - (maximpactistep) for i in range(samplesize)], threshold_indifference=0).
   • Set step and n for the 100 seeds run:
     • n: vary the n from 0 to 9 for each run to get the results for all 10 concern thresholds as required by Results_visualization.
   • Thresholds:
     • threshold_indifference: Defines the value below which impacts discourage recycling and selling (reuse).
     • threshold_concern: Defines the value above which impacts encourage recycling and selling (reuse).
5. Run the model: The model produces one CSV file per run (model results and agent results). Open batch_visualize.ipynb, adapt the USE trigger, and run the script to produce figures where impacts are compared to the percentage of landfill and circular EoL pathways.

Changes and Documentation

For detailed documentation about the new scripts and the changes in the old ones, please refer to the DOCUMENTATION.md file.

License

Our model is available under the EUPL license.

The source code for the ABM model was retrieved from https://github.com/NREL/ABSiCE and is used in accordance with the BSD 3-Clause License. Copyright (c) 2020, Alliance for Sustainable Energy, LLC.

The LCA model code is not open source. However, the metamodel and the stochastic inputs and outputs are available in the metamodel folder.