Optimize Energy Assets with energy-py-linear

Optimizing energy assets is critical for maximizing profit and minimizing carbon emissions - but most energy asset operators rely on simple heuristics or manual dispatch decisions.

Linear programming offers a better solution - it can find optimal dispatch schedules for batteries, electric vehicles, combined heat and power generators, and renewable assets.

This post introduces energy-py-linear - a Python library that makes it easy to optimize energy assets using mixed-integer linear programming (MILP).

You can find the source code at ADGEfficiency/energy-py-linear and the documentation at energypylinear.adgefficiency.com.

Why Linear Programming for Energy?

Linear programming is the right tool for many energy optimization problems - energy systems often have linear characteristics that make them well-suited for linear solvers.

Advantages:

• Deterministic: No randomness like gradient descent, same inputs always produce same outputs
• Fast: Linear programs are solved efficiently, feasible solutions exist on constraint boundaries
• Proven: Widely used in industry for energy dispatch and scheduling
• Interpretable: Easy to understand why the optimizer made specific decisions

What Can energypylinear Do?

energypylinear optimizes the dispatch of various energy assets:

• Batteries: Electricity storage with configurable power, capacity and round-trip efficiency
• Electric vehicles: Smart charging with flexible connection schedules
• CHP generators: Combined heat and power with gas consumption and heat recovery
• Heat pumps: Electrically-driven thermal energy conversion
• Renewable generators: Solar and wind with optional curtailment
• Boilers: Thermal generation from fuel combustion

Assets can be optimized individually or together as a site, with customizable objectives (profit, carbon, or custom) and constraints.

Installation

Requires Python 3.11 or 3.12:

$ pip install energypylinear

Example: Battery Arbitrage

The simplest use case is optimizing a single battery for price arbitrage:

import energypylinear as epl

asset = epl.Battery(
    power_mw=2.0,
    capacity_mwh=4.0,
    efficiency_pct=0.9,
    electricity_prices=[100.0, 50, 200, -100, 0, 200, 100, -100],
)

simulation = asset.optimize()

The battery will charge when prices are low (or negative) and discharge when prices are high.

Example: Multi-Asset Site Optimization

The real power of energypylinear comes from optimizing multiple assets together:

import energypylinear as epl

assets = [
    epl.Battery(power_mw=2.0, capacity_mwh=4.0),
    epl.CHP(
        electric_power_max_mw=100, electric_power_min_mw=30, electric_efficiency_pct=0.4
    ),
    epl.Boiler(),
    epl.Valve(),
]

site = epl.Site(
    assets=assets,
    electricity_prices=[100, 50, 200, -100, 0],
    high_temperature_load_mwh=[105, 110, 120, 110, 105],
    low_temperature_load_mwh=[105, 110, 120, 110, 105],
)

simulation = site.optimize()

The optimizer coordinates all assets to minimize cost while meeting thermal loads.

Example: Carbon-Optimized Renewable Site

energypylinear can optimize for carbon emissions instead of profit:

import energypylinear as epl

assets = [
    epl.Battery(power_mw=10, capacity_mwh=20, efficiency_pct=0.9),
    epl.RenewableGenerator(
        electric_generation_mwh=[10, 20, 30, 20, 10],
        electric_generation_lower_bound_pct=0.5,
        name="solar",
    ),
]

site = epl.Site(
    assets=assets,
    electricity_carbon_intensities=[0.5, -0.5, 0.5, 0.5, -0.5],
    export_limit_mw=25,
)

simulation = site.optimize(objective="carbon")

The results show how the battery and solar generation coordinate to minimize carbon:

   site-electricity_carbon_intensities  site-export_limit_mw  site-export_power_mwh  solar-electric_generation_used_mwh  battery-electric_charge_mwh  battery-electric_discharge_mwh
0                                  0.5                    25                  10.00                                10.0                        0.00                          0.00
1                                 -0.5                    25                  25.00                                20.0                        5.00                          0.00
2                                  0.5                    25                  25.00                                30.0                        0.00                          5.00
3                                  0.5                    25                  15.00                                20.0                        0.00                          5.00
4                                 -0.5                    25                  10.00                                10.0                        0.00                          0.00

Key observations:

• Interval 1: Negative carbon intensity, battery charges to enable maximum export
• Intervals 2-3: Positive carbon intensity, battery discharges to maximize export
• Export limit: 25 MW limit is respected throughout

Advanced Features

energypylinear supports advanced use cases:

• Custom constraints: Add your own linear constraints to the model
• Custom objectives: Define alternative objective functions beyond cost or carbon
• Network charges: Model complex electricity tariffs with time-of-use pricing
• Battery degradation: Account for capacity fade over time
• Dispatch forecasts: Optimize future dispatch based on forecasted prices

When to Use energypylinear

energypylinear is ideal for:

• Business case modeling: Evaluating investment in energy assets
• Day-ahead dispatch: Optimizing assets based on day-ahead price forecasts
• Carbon analysis: Understanding carbon impact of different dispatch strategies
• Multi-asset coordination: Finding optimal schedules across multiple assets

energypylinear is not suitable for:

• Real-time control: Too slow for sub-second dispatch decisions
• Nonlinear systems: Only models linear relationships
• Very large systems: Performance degrades with thousands of intervals or assets

Summary

energypylinear makes it easy to optimize energy assets using mixed-integer linear programming.

Key features:

• Multiple asset types: Batteries, EVs, CHP, heat pumps, renewables, boilers
• Flexible objectives: Optimize for profit, carbon, or custom objectives
• Site optimization: Coordinate multiple assets together
• Customizable: Add custom constraints and objectives
• Production-ready: Type-safe, well-tested, documented

Learn more:

• Documentation: energypylinear.adgefficiency.com
• Source code: github.com/ADGEfficiency/energy-py-linear
• Installation: pip install energypylinear

Thanks for reading!