Rationalising data collection for supporting decision making in building energy systems using value of information analysis

This repository supports the article 'Rationalising data collection for supporting decision making in building energy systems using value of information analysis', published by the Journal of Building Performance Simulation on 5th Nov 2024. It is available online at https://www.tandfonline.com/doi/full/10.1080/19401493.2024.2423827. It provides the code used to perform the Value of Information (VoI) calculations for the example problems presented in the paper.

Technical Requirements

Use of this codebase requires Python 3.9 or later. A suitable environment can be initialised using conda as follows:

conda create --name myenv python>=3.9
conda activate myenv
pip install -r requirements.txt
conda install -c conda-forge cmdstanpy

Codebase Structure

This repository contains:

• Three scripts for performing the VoI calculations for each example problem:
  1. building_ventilation.py
  2. ASHP_maintenance.py
  3. GSHP_design.py
• voi.py, a generic implementation of the EVPI & EVII calculations for a general one-stage Bayesian decision problem (including accelerated versions).
• The stan_models directory, containing Stan implementations of the probabilistic models used in the GSHP design example.
• The data directory, containing input data and cached utility evaluations for the GHSP design example.
• The models directory, containing an energy simulation model for the example GSHP system.
• The results directory, containing the numerical results of the sampling and evaluations used to perform the VoI calculations.
• The plots directory, containing code to produce the influence diagram figures for each example problem.
• The trial_scipts directory, containing scripts used for development and checking.
• A helpful caching function wrapper in utils/caching.py