Setting Up Brightway¶

optimex builds on the Brightway LCA framework. Before running an optimization, you need to set up a Brightway project with the required databases: a biosphere database for elementary flows, one or more background databases representing supply chains at different points in time (see Theory: Background Evolution), and a foreground database with the candidate processes to optimize.

Project Structure¶

An optimex project requires three types of databases:

Brightway Project
├── biosphere3          # Elementary flows (emissions, resources)
├── Background DBs      # e.g., db_2020, db_2030 (ecoinvent variants)
└── foreground          # Your processes to optimize (must be named "foreground")

Creating a Project¶

import bw2data as bd

# Create or switch to a project
bd.projects.set_current("my_optimex_project")

Biosphere Database¶

The biosphere database contains elementary flows (emissions to air/water/soil, resource extractions). If you're using ecoinvent, this is imported automatically. For custom setups:

biosphere_data = {
    ("biosphere3", "CO2"): {
        "type": "emission",
        "name": "carbon dioxide",
        "categories": ("air",),
        "CAS number": "000124-38-9",  # Helps with dynamic characterization
    },
    ("biosphere3", "CH4"): {
        "type": "emission",
        "name": "methane, fossil",
        "categories": ("air",),
        "CAS number": "000074-82-8",
    },
}

bd.Database("biosphere3").write(biosphere_data)

Using ecoinvent

If you import ecoinvent using bw2io, the biosphere database is created automatically with all necessary flows and CAS numbers.

Background Databases¶

Background databases represent the supply chain processes that support your foreground system. In optimex, you can have multiple background databases at different time points to model technological evolution.

Single Background Database¶

from datetime import datetime

# Create background database
bg_2020 = bd.Database("db_2020")
bg_2020.write({
    ("db_2020", "electricity"): {
        "name": "electricity production",
        "location": "GLO",
        "reference product": "electricity",
        "exchanges": [
            {"amount": 1, "type": "production", "input": ("db_2020", "electricity")},
            {"amount": 0.5, "type": "biosphere", "input": ("biosphere3", "CO2")},
        ],
    },
})

# Set representative time (required for optimex)
bg_2020.metadata["representative_time"] = datetime(2020, 1, 1).isoformat()
bg_2020.register()

Multiple Time-Specific Databases¶

To model technology improvement over time, create multiple versions:

# Database for 2020
bg_2020 = bd.Database("db_2020")
bg_2020.write({...})  # Current technology
bg_2020.metadata["representative_time"] = datetime(2020, 1, 1).isoformat()
bg_2020.register()

# Database for 2030 (cleaner technology)
bg_2030 = bd.Database("db_2030")
bg_2030.write({...})  # Improved technology with lower emissions
bg_2030.metadata["representative_time"] = datetime(2030, 1, 1).isoformat()
bg_2030.register()

# Database for 2040
bg_2040 = bd.Database("db_2040")
bg_2040.write({...})  # Even cleaner
bg_2040.metadata["representative_time"] = datetime(2040, 1, 1).isoformat()
bg_2040.register()

optimex automatically interpolates between these databases based on when exchanges occur.

Using premise

The premise package can automatically generate future versions of ecoinvent based on IAM scenarios. This is the recommended approach for realistic prospective LCA.

Foreground Database¶

The foreground database contains the processes you want to optimize. It must be named "foreground".

See Foreground Modeling for detailed guidance on setting up foreground processes with temporal distributions.

fg = bd.Database("foreground")
fg.write({
    ("foreground", "process_A"): {...},
    ("foreground", "process_B"): {...},
})
fg.register()

Characterization Methods¶

Define impact assessment methods for your elementary flows:

# Simple GWP method
bd.Method(("GWP", "100 years")).write([
    (("biosphere3", "CO2"), 1),
    (("biosphere3", "CH4"), 28),
])

# You can define multiple methods
bd.Method(("Water use", "m3")).write([
    (("biosphere3", "water"), 1),
])

Using existing methods

If you import ecoinvent with bw2io, standard methods like IPCC GWP and ReCiPe are available automatically.

Complete Setup Example¶

import bw2data as bd
from datetime import datetime

# 1. Create project
bd.projects.set_current("my_optimex_project")

# 2. Biosphere (if not using ecoinvent)
bd.Database("biosphere3").write({
    ("biosphere3", "CO2"): {"type": "emission", "name": "CO2"},
})

# 3. Background databases with time metadata
for year in [2020, 2030]:
    db = bd.Database(f"db_{year}")
    db.write({
        (f"db_{year}", "electricity"): {
            "name": "electricity",
            "exchanges": [
                {"amount": 1, "type": "production", "input": (f"db_{year}", "electricity")},
                {"amount": 0.5 * (1 - (year-2020)*0.01), "type": "biosphere", "input": ("biosphere3", "CO2")},
            ],
        },
    })
    db.metadata["representative_time"] = datetime(year, 1, 1).isoformat()
    db.register()

# 4. Foreground (see Foreground Modeling guide)
fg = bd.Database("foreground")
fg.write({...})  # Your processes
fg.register()

# 5. Characterization method
bd.Method(("GWP", "example")).write([
    (("biosphere3", "CO2"), 1),
])

Verifying Your Setup¶

Check that all databases are correctly configured:

# List all databases
print(bd.databases)

# Check background database metadata
for db_name in bd.databases:
    db = bd.Database(db_name)
    if "representative_time" in db.metadata:
        print(f"{db_name}: {db.metadata['representative_time']}")

# Verify foreground exists
assert "foreground" in bd.databases, "Foreground database required!"

Next Steps¶

Foreground Modeling: Set up foreground processes with temporal distributions, operation flags, and vintage parameters
Optimization Setup: Configure demand, characterization, and run the optimization