Skip to content

Set up an Antares-Xpansion study

Overview

The Antares-Xpansion package is based on the Antares software and its data format.

Antares-Xpansion is based on an already existing Antares study. Some of the capacities of this study, usually fixed as input in the Antares paradigm, will be optimized by the investment optimization module of the Antares-Xpansion package.

In order to run the investment optimization module, the Antares dataset must be enriched with - at least - two new files:

  • A candidates.ini file which contains the definition of investment candidates (which capacities of the Antares study are expandable? at what cost? with what limits? and so on),

  • A settings.ini file which contains the settings of the Antares-Xpansion algorithm.

These two files must be located in the user/expansion/ directory of the Antares study (see Figure 1). The data they contain are neither visible nor modifiable in the Antares man-machine interface. These two files must therefore be handcrafted.

antares-study
└─── input
└─── layers
└─── logs
└─── output
└─── settings
└─── user
│   └───expansion
│       │   candidates.ini
│       │   settings.ini
│       │   ...
Figure 1 - Structure of an Antares study folder ready for an Antares-Xpansion optimization

Depending on the type of investment considered in the study, some modifications of the Antares study are necessary before filling the candidates.ini and settings.ini files.

Prepare the Antares study

The first step to set up an Antares-Xpansion study consists in defining investment candidates. Candidate capacities for investment are necessarily links from an Antares study. Investment candidates can be generation assets, or even flexibilities, by adopting a virtual node logic as described below.

(a) (b) (c)

Figure 3 - Configuration of the Antares study for an investment in (a) transmission capacity (new line or grid reinforcement), (b) generation units and (c) storage.

The following sections describe the Antares modelling for different types of investment:

Investment in transmission capacity between two areas

The Antares link candidate for investment, shown in red in Figure 3 (a), is directly the interconnection for which the interest in increasing capacity is being studied.

  • In the case of the construction of a new line, a link must be added in the Antares study between the two involved areas.

  • In the case of a grid reinforcement between two already interconnected areas, the link between these two areas already exists in the Antares study. The parameter already-installed-capacity will be used in the candidates.ini file to specify the capacity of the grid that already exists between the two zones. In this way, Antares-Xpansion will assess the economic interest of increasing this capacity beyond what is already installed.

Investment in thermal generation capacity

Suppose that the thermal generation capacity subject to expansion is physically located in area1 for the example in Figure 3 (b). As Antares-Xpansion only performs investment through links of the study, the modelling trick consists in creating a virtual node, here invest_semibase, connected to the physical node area1. The Antares link of the investment candidate is the link between these two nodes.

The generation unit, candidate for investment, must be defined by a thermal cluster with the following technical and economic parameters:

  • It is located in the virtual node invest_semibase.

  • Its market bid is equal to its marginal cost, which is the variable operating cost (in €/MWh) of the generation unit.

  • It has an hourly availability time series that is always higher than the potential of the candidate, where by potential we mean the capacity correponding to the maximum investment that is allowed by the user (i.e. max-investment or max-units \( \times\) unit-size):

    • If the hourly availability time series of thermal generation are "ready-made” in Antares, then the values of the time series must be filled in such a way that they are always higher than the candidate's potential.

    • If the times series of thermal generation are “stochastic” i.e. generated by Antares, then the parameters for the generation of the series must be defined so that the availability is always higher than the potential (number of units \( \times\) nominal capacity \( >\) potential, no outages rate).

Other cluster parameters (pmin, start-up costs, etc.) can also be defined. However, they will only be taken into account by Antares-Xpansion if the unit-commitment type is set to expansion_accurate.

Investment in renewable generation capacity

Similarly as for thermal generation, the renewable generation capacity subject to expansion, physically located for the example in Figure 3 (b) in area1, must be moved to a virtual node connected to the physical node area1. The Antares link for the investment candidate is the link between these two nodes.

For the type of renewable production of interest (wind or solar), a corresponding production time-series (Antares wind or solar tab) must be defined in the virtual node. The production time-series must be deterministic, constant, and higher than the potential (max-investment) of the candidate. The parameters [in]direct-link-profile from the candidates.ini file will then be used to define the hourly load factor.

Investment in flexibility

The modeling of flexibilities, such as pumped storage, is generally based in Antares on a set of virtual nodes/links and coupling constraints. To make flexibility an investment candidate, a link must be identified in the Antares modelling whose transmission capacity corresponds to the capacity of the flexibility (e.g. its maximum power or the size of a stock for example).

In the case of pumped storage in Figure 3 (c), the capacity of the pumped storage (equal to its pumping and turbining capacity) is defined by the maximum possible flow on the link between area1 and hub: the investment in the flexibility will be characterized by this link. The classical binding constraints must be added in the Antares simulation to represent the storage: for example a negative ROW Balance in psp-in, positive ROW Balance in psp-out and the following constraint:

Decommissioning decisions for thermal capacities

With Antares-Xpansion, it is possible to consider decommissioning decisions, the corresponding assets are referred as decommissioning candidates. The difference between investment candidates and decommissioning candidates lies in the fixed-cost annuity.

Fixed-cost annuity for investment candidates

The annuity of an investment candidate includes the sum of:

  • Annualized investment costs,
  • Fixed annual operation and maintenance costs.

In this configuration, Antares-Xpansion makes an economic choice by comparing the sum of these costs and the reduction in variable operating costs (mainly fuel costs and penalties associated with loss of load) due to the new investment.

Fixed-cost annuity for decommissioning candidates

The annuity for decommissioning candidates only includes the fixed annual operation and maintenance costs. There is indeed no investment cost, since the decision consists only in choosing whether to maintain operation with the associated maintenance costs.

In this configuration, Antares-Xpansion makes an economic choice by comparing the operation and maintenance costs of a generation or transmission asset and the savings induced on the variable costs of power system operation thanks to this asset.

The annualized investment costs are in this case considered stranded and are not taken into account in this economic choice. The potential of this type of candidate (i.e. its max-investment or max-units \( \times\) unit-size) corresponds to its decommissionable capacity, or in other words, the candidate's already installed capacity that could be shut down if it is no longer profitable for the power system.

Antares-Xpansion is not able to decommission generation units that are installed in the Antares study (i.e. located in "physical nodes"). However, we can use a modelling of decommissioning candidates with the same virtual node logic as the investment in thermal generation capacity. Decommissioning candidates are existing physical facilities that should be moved to a virtual node.

For example, suppose that we aim at taking a decommissioning decision for thermal generation capacities that are physically located in area1 of Figure 3 (b). To be considered decommissioning candidates, these generation units must be moved to a virtual node (invest_semibase in Figure 3 (b)) with an hourly availability time series higher than their potential.

The decommissioning decision is made in Antares-Xpansion through the capacity of the link between area1 and invest_semibase. Thus, the capacity invested by Antares-Xpansion on the link corresponds to the capacity that is not decommissioned.

Details on how to fill in the file candidates.ini for decommissioning candidates are given in the next part.

In the four aforementioned cases, the link used to define investment candidates (in red in Figure 3):

  • must have the parameter transmission capacities = use transmission capacities, and not set to null or set to infinite,

  • may have a hurdle cost, which will then be well taken into account in the economic optimization of Antares-Xpansion,

  • may be subject to binding constraints - provided that the Antares version used is at least v6.1.3 - which will be well taken into account in the simulations of system operation. These constraints can possibly be constructed by the Kirchhoff constraint generator and the information given in the impedances, loop flow and phase shift columns of the link.

The direct and indirect transmission capacities of the link will be modified by Antares-Xpansion. The values initially entered in the Trans. Capacity Direct and Trans. Capacity Indirect columns do not matter since they will be overwritten when the expansion problem is solved. Note that the capacities of existing structures must be filled in with the already-installed-capacity parameter in the candidates.ini file and not in the definition of the links in the Antares study.