Skip to content

Features

PyPSA is a flexible framework for modelling and optimising modern energy systems. It supports high spatial, temporal, and sectoral resolution, from short-term dispatch to long-term planning. The following feature set outlines its key features:

Optimisation Functionalities

  • Economic Dispatch (ED): Models short-term market-based dispatch including unit commitment (either with integer variables as MILP or in a relaxed approximation as LP), renewable availability, short-duration and seasonal storage including hydro reservoirs with inflow and spillage dynamics, elastic demands, load shedding and conversion between energy carriers, using either perfect operational foresight or rolling horizon time resolution.

  • Linear Optimal Power Flow (LOPF): Extends economic dispatch to determine the least-cost dispatch while respecting network constraints in meshed AC-DC networks, using a linearised representation of power flow (KVL, KCL) with optional loss approximations.

  • Security-Constrained LOPF (SCLOPF): Extends LOPF by accounting for line outage contingencies to ensure system reliability under \(N-1\) conditions.

  • Capacity Expansion Planning (CEP): Supports least-cost long-term system planning with investment decisions for generation, storage, conversion, and transmission infrastructure. Handles both single and multiple investment periods. Continuous and discrete investments are supported.

  • Pathway Planning: Supports co-optimisation of multiple investment periods to plan energy system transitions over time with perfect planning foresight.

  • Rolling-Horizon Optimisation: Enables sequential optimisation of operation with myopic foresight, allowing for dynamic information updates and breaking down large problems into manageable time slices.

  • Stochastic Optimisation: Implements two-stage stochastic programming framework with scenario-weighted uncertain inputs, with investments as first-stage decisions and dispatch as recourse decisions.

  • Modelling-to-Generate-Alternatives (MGA): Explores near-optimal decision spaces to provide insight into the range of feasible system configurations with similar costs.

  • Policy Constraints: Built-in support for policy constraints such as CO2 emission limits and pricing, subsidies, resource limits, expansion limits, and growth limits. Extendable by custom constraints.

  • Custom Constraints: Users can impose own objectives, variables and constraints, such as policy constraints or technical requirements, using Linopy.

  • Solver Flexibility: Supports a wide range of LP, MILP, and QP solvers from open-source solutions (e.g. HiGHS, SCIP) to commercial products (e.g. Gurobi, COPT).

Use Cases and Grid Modelling

  • Sector-Coupling: Modelling integrated energy systems with multiple energy carriers (electricity, heat, hydrogen, etc.) and conversion between them. Flexible representation of technologies such as heat pumps, electrolysers, battery electric vehicles (BEVs), direct air capture (DAC), and synthetic fuels production.

  • Diverse Applications: Supports a wide range of energy system analyses for strategic decision support. Applications include techno-economic assessment of technologies, capacity expansion, transmission planning, market design, sector-coupling, integration of variable renewables such as wind and solar, flexibility needs and resource adequacy assessments, network congestion analysis, battery scheduling, electricity trading, planning of decarbonisation pathways, hydrogen infrastructure planning, electrolyser siting and operation, resilience to extreme weather, bidding zone configurations, and the design of islanded systems for remote renewable fuel production.

  • Standard Grid Components: Includes standard types for lines and transformers from pandapower.

  • Static Power Flow Analysis: Computes both full non-linear and linearised load flows for meshed AC and DC grids using Newton-Raphson method with optional distributed slack and shunt compensation.

Architecture and Performance

  • Modular Design: Clean separation between data and modelling code enables flexible scenario development.

  • Resolution Control: Offers flexible control over temporal, spatial, and sectoral scope and detail.

  • Spatial Clustering: Can reduce model size for large networks via spatial aggregation strategies.

  • Data Backbone: Uses pandas for data handling and linopy for optimisation and interfacing with solvers.

  • Performance: Using linopy, the code is designed to scale well with high resolution networks, minimising memory usage and time spent outside the solver.

  • Scalability: Handles models ranging from small conceptual prototypes to continent-scale high-resolution systems solved on high-performance compute clusters.

Analysis and Usability

  • Shadow Prices: Outputs dual values such as nodal clearing prices (LMPs), storage water values, scarcity and CO2 prices.

  • Statistics: Built-in tools for summarising and visualising results, such as energy balances, capacities, costs, market values, curtailment, component revenues.

  • Visualisations: Built-in tools for plotting statistics, time series data and spatial distributions of line loadings and nodal dispatch decisions.

  • Documentation: Comprehensive user guide, API reference, and plenty of examples are available.

  • Community Support: Active community on GitHub and Discord for user support and development discussions.

  • MIT License: Fully open-source and free for commercial and academic use.

Illustrations

Interactive network visualization of SciGRID example network clustered by federal state, showing transmission capacities, flow directions, electricity supply (upper pie charts) and demand (lower pie charts), as well as average nodal prices (color scale):

Interactive area plot of electricity balance time series in a highly-renewable sector-coupled example network, showing temporal generation (positive values) and consumption (negative values) of different technologies: