e-Highway 2050: Grid reinforcements design for long term transmission planning in Europe

Challenges
- How to identify the grid bottlenecks caused by physical power flows exchanged between areas in Europe at the 2050 time horizon?
- Which grid reinforcements could tackle these constraints for a given energy scenario for the pan-European transmission system at 2050?

Background and assumptions 
The analysis focusing on the system in 2050 is based on a scenario approach projecting the outlook of the energy system in Europe in 2050 and on the already known grid reinforcements scheduled until 2030 which constitute the “starting grid” of the approach, common for all scenarios:

- The five scenarios for the pan-European transmission system at 2050 have been described here and Europe has been clustered into about 100 different clusters modelling a simplified European transmission system (assuming 1 cluster = 1 node)

Figure 1. Countries and clusters including North Sea and North Africa (left) and Equivalent transmission network (right)

- The quantification of the five scenarios (meaning that demand, storage, exchange and generation have been defined up to the cluster level) was carried out thanks to a dedicated approach.
- The starting grid was set assuming that the transmission network existing today will still be in operation by 2050 and that the transmission network developments foreseen by the TYNDP 2014 until 2030 (1) will all be completed.

All these preparatory steps were needed to perform system simulations modelling the behavior of the whole European power system with robustness considering both the starting grid as initial condition and the evolutions of generation and demand over the period. Two successive steps are implemented. First the identification of network constraints by measuring difference between a “copper plate” grid and a grid with limited capacities. Then, in a second step, Transmission Requirements are defined representing the needed increase of capacity between two clusters. They are progressively defined by an iterative approach with the goal to minimize the effects of grid limitations at least costs.

1. System simulations
The system simulations optimize the dispatch of generation in terms of cost for each hour of the year, taking into account the topology and characteristics (impedances and grid transfer capacities) of the “starting grid”. The optimization process identifies the required inefficient generation shift allowing electricity flows to remain within their limits. The operating cost of the pan-European power system can be estimated for the “starting grid” and for any modified grid architecture.

The simulations are performed with the ANTARES (2) tool:
- The optimisation aims to minimize the overall cost of the system taking into account grid characteristics. The European system is optimized in one shot and a perfect market is assumed.
- Probabilistic simulations of 99 possible years are performed thanks to a Monte-Carlo approach (3) to address the variability inherent to a high share of RES
- The time step resolution is one hour and simulations cover a period of one year.

2. The methodology aiming at detecting the constraints
This first phase identifies the most critical issues for the grid in both time and space dimensions while taking into account the intrinsic uncertainty of some variables (e.g. renewables).

Network constraints effects are measured by the difference between two simulations: The “copperplate” situation - in which network capacities are set to infinite – and a mode “with grid constraints” - in which capacities are limited to the “starting grid”(4) at first, then to the “starting grid” with Transmission Requirements of the reinforcement process.

The “copperplate” simulation gives the upper limit of what could be achieved by grid reinforcements to ensure system security and optimize operating costs. On the contrary, the “starting grid” simulation gives the lowest level of system security that can be achieved with the 2030 transmission network status after implementation of the demand and generation levels as expected for 2050.

To address the tractability of a huge amount of data (8760 hours x 99 MC years for each of the about 100 clusters and each link (>200)), a progressive approach in both time and space dimensions is implemented with an in-depth analysis of three indicators for each scenario at different time and geographical scales:
- ENS (Energy not served or unsupplied energy) represents the volumes of energy not served due to network limitations
- Extra spillage or delta spillage depicts the must run energy (e.g. Renewables) that cannot be consumed due to the congested grid
- Thermal redispatch which can be positive or negative: a positive redispatch in a given cluster meaning that local thermal generation is increased to secure the load. This (more expensive) generation substitutes RES energy and competitive thermal generation available in other clusters but that cannot be relieved due to network limitations. Conversely, a negative redispatch means that thermal generation that is optimized in the copper plate simulation is reduced due to grid constraints (typically it occurs in countries/clusters with capacities in competitive technologies, with excess of energy).

3. The progressive grid development towards the architecture at 2050
An iterative process is implemented which stops when indicators are stabilized. Once all the reinforcements are identified, they are transposed into possible technologies for a cost / benefit assessment of the resulting architecture (the complete set of reinforcements for the pan-European space).

The loop is detailed in the figure below: based on constraints analysis, transmissions requirements are proposed and tested.
 

Fig. 2: Iterative process to define the reinforcements

The constraints analysis of characteristic weeks enable to suggest reinforcements to connect areas in surplus to areas in deficit. Their sizes are set based on the identification of synchronous volumes at stake. The possibility to collect or distribute energy along the path is considered.


Figure 3. Example of map with average hourly values on the 99 Monte Carlo years
in a characteristic week and reinforcements

The reinforcements are not tested one by one or all at once: at each iteration of the process, a set of reinforcements constitutes a “step” to be further assessed.

- The annual benefit of the “step” is calculated as the sum of generation costs savings and ENS costs reduction. Within the project, ENS costs are estimated on the level of 10000 Euro/MWh for the whole Europe (a cost of 1000 Euro/MWh is also considered in a sensitivity analysis).
- The cost of the tested set of reinforcements is then roughly assessed considering costs of DC cables, as the most expensive case. It is compared to the benefits to verify that the investments are profitable whatever the technology as a “no regret investments”. Based on the remaining constraints in the characteristic weeks, a new iteration of reinforcement is then defined. If there are no more significant issues (which means only small and spread volumes of ENS, spillage and redispatch remain), the iterative process stops.

The final grid proposal is made of all the transmission requirements (reinforcements) defined in the different steps and defines the final architecture.

4. Technology and cost assessment
Transmission Requirements consist in specification of needs and are not related to any specific technology: they are however transposed into possible technologies to ensure the technology availability and to better assess the cost of the final architecture.

Indeed since the selection of technologies in 2050 will be highly impacted by the level of public acceptance towards new lines, three strategies are considered to encompass a wide range of possible costs: from a “Status Quo” strategy (5) to a “New Grid Acceptance” strategy (6), an intermediate strategy being the “Re-Use of Corridors” in which the public accepts new overhead lines as long as they are close to existing ones.

It should be noted that a given transmission requirement can be realized through many parallel lines, especially due to the capacity limit of a single line and the N-1 robustness. These different lines could follow different routes between the clusters and be connected to different substations, but this perspective remains beyond the scope of this study.

5. Cost/benefit analyses
A simplified analysis is carried out as a final step here, while a more comprehensive toolbox for the cost and benefit analysis applicable to a reinforcement project is proposed here. For the three strategies, the investment annuity (7) of the final architecture is compared to the annual benefit of it, to measure the level of profitability.

This assessment of the profitability encompasses the whole architecture and does not evaluate profitability of partial packages of reinforcements (the profitability of each reinforcement depends strongly on the other reinforcements already in place, thus it is difficult to assess the profitability of a single project).

6. Resulting grid architectures at 2050
New transmission reinforcements to support each of the five identified scenarios confirm the predominance of “North to South” corridors. An overview of architectures is proposed and discussed here.

References
[1]    B. H. Bakken, M. Paun, R. Pestana, G. Sanchis, “e-Highway2050: A Modular Development Plan on Pan- European Electricity Highways System for 2050”, Cigre Lisbon, April 2013
[2]    G. Sanchis, RTE et alia, “A methodology for the development of the pan-European Electricity Highways System for 2050”, CIGRE Paris, August 2014  
[3]    Thomas Anderski, Amprion et alia, deliverable D2.3 System simulations analysis and overlay-grid development – Digest.
[4]    T. Anderski, Amprion; F. Careri, RSE; N. Grisey, RTE; G.Migliavacca, RSE; D. Orlic, EKC; G. Sanchis, RTE. e-Highway2050: a research project analysing very long term investment needs for the pan-European transmission system. Cigre Paris. Submitted to Cigre Paris, August 2015
[5]    http://www.e-highway2050.eu
[6]    Thomas Anderski, Amprion et alia, deliverable D2.4 Contingency Analyses of Grid Architectures and Corrective Measurements, to be published in 2015
[7]    ENTSO-E, “Ten-Year Network Development Plan (TYNDP)”, www.entsoe.eu/major-projects/ten-year-network-development-plan/, 2014

Notes
(1) which includes for example major North-South HVDC corridors in Germany
(2) Antares is a sequential Monte-Carlo system simulator developed by RTE. More details can be found in: M. Doquet, C. Fourment, J.M. Roudergues “Generation & Transmission Adequacy of Large Interconnected Power Systems : A contribution to the renewal of Monte-Carlo approaches”, PowerTech2011, IEEE Trondheim
(3) 11 different solar & wind “years” combined with 3 different demand “years” –based on temperatures) and 3 types of hydrologies -dry, average, wet “years”
(4) Starting grid 2030 based on the actual grid and the TYNDP2014 enlargements 
(5)the public opposition against new infrastructure prevents any new OHLs. Only refurbishment of existing lines or new DC cables can be implemented
(6) the public accepts new OHL and also the development of new corridors. DC cables are also possible but OHL are preferred when possible due to their lower costs
(7) Investment annuity is calculated as discounted annual costs respecting economic lifetime of the equipment defined in WP3 and discount rate of 5%.