Overview#
This chapter aims to give a broad overview of the use of PyPSA-Eur in the context of the Energie Studie 2050+ for the Flemish Energy and Climate Agency (VEKA, Belgium).
See also
Web interface to explore the data online
High level description#
PyPSA-Eur is an open model dataset of the European energy system at the transmission network level that covers the full ENTSO-E area. Written in Python, it optimizes the energy system with high temporal and spatial resolution for a large spectrum of technologies and energy vectors. It covers demand and supply for all energy sectors.
General description#
PyPSA-Eur optimizes energy systems with
a temporal granularity up to 1 hour (which represents a maximum of 8760 time steps each year) ;
a spatial granularity up to ENTSO-E transmission lines (which represents a maximum of 8800 electrical buses) ;
a technology list reaching up to 80 different technologies (see Technological assumptions).
PyPSA-Eur allows connecting different energy vectors (referred to as energy carriers in PyPSA-Eur) and technologies using them, hence allowing to tackle inter-sectorial and inter-carriers connections.
The system is optimized over the energy system cost (including OPEX and annualized CAPEX), for different consecutive planning horizons. The optimization is done over the sizing and dispatch of energy production and storage units and of transmission infrastructures to meet an inelastic demand for each of the time frames and geographical locations considered.
The cost optimization of the energy system for a given year is performed under different constraints, including an annual CO2 emission constraint for the given year and renewable potentials.
Optimization details#
For each planning horizon, PyPSA-Eur provides the cost optimal energy dispatch of each asset at each node and for each time frame given an hourly demand at each node of each energy carrier, as well as the cost optimal sizing of production, storage and transmission units needed for such dispatch (see Power System Optimization)
The optimization hence minimizes the annual total cost of the energy system for each planning horizon, which is defined as :
where :
\(CAPEX_{n,s}\) is the annualized investment cost of units at node n for generator or storage asset s
\(CAPEX_{l}\) is the annualized investment cost of infrastructure on line l
\(OPEX_{n,s,t}\) is the operational cost of units at node n, for generator or storage asset s at time frame t
\(w_{t}\) is the weighting of time t in the objective function (e.g. multiple hours)
If myopic optimization is configured, the solver does not optimize the energy system over a continuous trajectory but rather planning horizon by planning horizon. For each planning horizon, the capacity installed in the previous planning horizon is taken into account and phased out assets are removed.
Given the spatial and temporal resolution used, the use of a commercial solver is required to compute the solution in a reasonable amount of time. Therefore, a Gurobi license, which is the standard choice in the PyPSA community, has been bought.
Optimization constraints#
The energy system is optimized under various constrains, among which :
A carbon budget constraint, representing how much CO2 can be emitted over the operation of the energy system for the considered planning horizons ;
A potential constraint, representing the maximal capacity that can be installed per technology and per node. Currently, renewables are the only technologies for which this potential constraint applies.
Main limitations#
The current study has been developed on two successive versions of PyPSA-Eur (v0.8.0 and v0.10.0) (PyPSA/pypsa-eur). Please refer to the release notes for further details on fixes and improvements since then. A limitations list is also maintained by the PyPSA-Eur team. During the development of this project, the following attention points have been identified :
For industry, the production of different materials per country is assumed to remain constant and no industry demand elasticity is included in the modelled. Hence, no Demand Side Management is modeled for the industry.
Energy carrier transmission/transport:
Hydrogen and electricity are by default regionalized (i.e. considered at for each nodes) and transmission infrastructure are modeled.
Methane and CO2 are by default regionalized without transport infrastructures. We recommend to include these infrastructures.
Solid biomass is by default modeled as a single equivalent EU node. The energy vector can be regionalized with or without transport between nodes. Transport of solid biomass between nodes does not include capital costs but only marginal costs and maximum nominal power between two nodes. The current configuration consider regionalized topology and transport.
Heat is regionalized but excluding district heating, heat is not transported.
Ammonia and oil are by default modeled as a single equivalent EU node, but can be regionalized (without transport infrastructure). We recommand to regionalize these carriers.
Coal, lignite and methanol are by default modeled as a single equivalent EU node, but can be regionalized. We recommand to keep the EU node.
Uranium is not regionalized and a single equivalent EU node is considered.
Some technologies are not yet supported by the model, including:
Geothermal energy for heat or electricity production ;
Industrial heat pumps ;
Low TRL storage technologies (Flow batteries, Compressed Air Energy Storage, gravitary storage system, Pumped Thermal Energy Storage, etc) ;
Exports are not taken into account in the current version of the model. The only way to model them is exogenous.
Hydrogen can be:
used as a feedstock/energy carrier for industry and as an energy carrier for different processes ;
produced and is not imported ;
transported in a gaseous form through pipelines ;
used as liquefied hydrogen or hydrogen for shipping demand.