The main challenge relates to the building of cost trajectories until 2050 for a selection of transmission technologies relevant for the power system and consistent with the defined eHighway2050 scenarios.
Background and assumptions
In the general context of the technology assessment of the most impacting technologies for the power system in the EU28 at the 2050 time horizon a particular focus was made on power transmission technologies and more specifically on their cost trajectories.
It is reminded that the scope of power system technologies covers the whole electricity value chain from generation and storage, transmission (passive and active transmission technologies) to demand and that the technology portfolio of power system technologies is defined upon selection criteria that are based on their impact on transmission networks with regard to planning issues by 2050.
Description of the result
The technology assessment of available and commercially mature power system technologies at 2050 is the result of a collective work under the management of Technofi and involving key European stakeholders of the electricity value chain (manufacturers, TSOs, academia). It consists first in the data gathering of technical performance data on commercially mature technologies at each time horizon and then in an appraisal of tentative costs for the next four decades, i.e. from today to 2050.
The database on transmission technologies includes AC, DC or hybrid interconnections or Power Electronics to better control flows over long distances. Transmission technologies have been organized into:
Figure 1: Portfolio of transmission technologies in the e-Highway2050 technology database
The construction of future cost trajectories for transmission equipment based on their today’s estimated cost is one of the three major issues for the construction of the database and is detailed below.
1. Proposed challenge and first observations
Estimating likely evolutions of costs of transmission equipment, beyond a short-term “grid planning” time horizon, remains a complex exercise. Several sources propose models for estimating costs of technologies according to their maturity. These learning curve or experience curve based models have been explored in-depth for generation and demand technologies. They predict cost dropping rates per time period according to the market penetration of the technologies. Scientific literature on the “experience curve” approach applied to power transmission is less abundant1.
Predicting costs of transmission technologies is difficult since several exogenous factors might significantly impact the forecasts, i.e. the prices of commodities such as copper (cost multiplied by a factor 3 in few years) or the price of oil. Furthermore, each transmission project is very specific and costs depend largely on the selected technologies and on local constraints (terrain, labor costs, social acceptance, etc.). When considering the cost structure of transmission project, one can observe that the variations in costs due to different initial conditions (terrain, etc.) can offset by an order of magnitude the uncertainties related to the forecast exercise. This means that a big effort should be spent on these initial conditions, i.e. attention must be paid to at least two key factors: the archetypal configuration (the precise description of the installed transmission system) and the geographical factor (terrain).
2. A systematic approach of deconstruction/reconstruction: key assumptions
It is assumed and verified that time evolution of cost trajectories can be modelled by a systematic breakdown of costs in five distinct components (equipment, installation, civil work, project management, authorizations and right of ways).
Evolution of each of these components can be approached via a simple evolution law with time constants that are inherent to the component. One could retain that three types of costs components could be distinguished with regard to evolution laws:
- cost components highly dependent on local constraints requiring a spatial analysis (terrain, country),
- cost components highly dependent on factors for which forecasts at a long-term time horizon remain difficult due to a disruptive event (external factor or disruptive technology),
- cost components for which evolution laws for the next decades could be built based upon basic assumptions under uncertainty margins.
A series of assumptions are then formulated to allow optimizing the trade-off between complexity and tractability. It is assumed that:
- The cost of a given archetype installed in a given installation context is known as of today
- No disruptive change in the macro-economic context (geopolitical instability, major economic crisis, no force majeure event),
- Evolution laws of rights of way and authorizations will depend on local constraints. This cost component, representing 4-10% of the overall costs, is excluded from the present analysis (i.e. no evolution law),
- Evolution laws of installation, civil works and project management will mainly depend on future evolutions of costs of energy and labor. Oil prices, labor and engineering cost type indices could be good proxies to capture their respective evolutions.
- There is a continuity in the technological evolution. The long-term trend for the future will thus result from the recent past trend and will take into account a classical technology learning curve2. Such an assumption has some limitations on the short-term-fluctuations that might create some bias but it is expected that these short-term fluctuations will be averaged when considering longer term periods.
- Evolution laws of equipment are thus driven by two complementary factors: the experience factor (i.e. a technology experience effect assuming no disruptive technology) to capture the ever progressing maturity of the industrial system. The second factor relates to the fact that transmission equipment (and especially lines and cables) includes raw materials (e.g. aluminum, copper), long-term trends of commodity prices has also to be taken into account.
3. Building the cost trajectories of transmission equipment until 2050
It is reminded that the technology database includes a limited number of technology archetypes, commonly built by manufacturers and TSOs and aiming to be sufficiently representative of transmission solutions commercially available and competitive over the period 2014-2050. This approach was adopted to reduce the complexity (cf. knowledge article e-Highway 2050: Management of uncertainties and data contextualization in the framework of the technology assessment of power system technologies expected to impact grid architecture studies at 2050,).
As a result, and for each considered transmission technology, cost evolutions can be estimated thanks to the aggregation of tentative forecasts of a series of indices at a given time horizon: commodity prices for energy and metals, labor and engineering costs as well as dropping rates (experience curve approach or other proxy).
If we exclude the local/zonal costs, the minimum number of components to be considered is five, namely LAB and OIL for Installation and Civil works components, ENG for project management component, and the two indices capturing the equipment component (i.e. EXP, METAL indices). This number could even be reduced to four distinct indices if the two last indices on experience and commodities are formulated as a synthetized index (under the form of dropping rates of cost) which reflects the cost reduction of the supplied (not installed) transmission equipment.
Figure 2: Approach to model cost trajectories based on cost breakdown and representative indices
The figure below details the simple evolution models assumed for each category of indices.
Figure 3: Breakdown of cost components in indices and type of model of time evolution for each category of indices
The aggregation is then direct for each time period for a given technology archetype. The figure below details
- The inputs: the initial CAPEX (1200 k€/km in the illustration below), the breakdown into the five components, the evolution laws of the end-indices (EXP, LAB, OIL, ENG), the uncertainty margins for each time period (±10% in 2030)
- The outputs: the CAPEX for each time period with an uncertainty level in min-max intervals.
Figure 4: Costs projection at 2050 of the AC OHL archetype, 400 kV, 4.3 GW, in rural plain
Finally, it should be added that if a variant of a given archetype has to be considered, it is proposed to resort to multipliers, (e.g. for other power, terrains, different number of conductors, different level of voltage, etc.).
Assessment of the methodology use and limitations
Forecasts at a long-term time horizon remain a tricky exercise with high degree of uncertainties. This is why an approach based on likely scenario and on a limited number of tentative archetype commonly built by manufacturers and TSOs was implemented to reduce the complexity.
Beyond the values and records included in the techno-economic database, one key value of the result is in its methodological dimension, e.g. it provides statements and assumptions in a modular way to rerun the various methodological blocks, e.g.:
- to recalculate future costs of transmission technologies based on different archetype or different breakdown of cost or different laws of evolutions of indices
- to re-contextualize data according to new 2050 scenarios
- to add non considered technologies beyond the current eHighway2050 technology portfolio.
The proposed approach could be easily reused:
- Within the eHighway2050 project to reassess cost of transmission equipment according to different input hypothesis
- Beyond the e-Highway2050 project time horizon and scope, in further grid planning studies at national or EU levels.
A future user of such methodology can adjust the simulations at different levels:
- use different initial cost conditions
- modify confidence levels through the uncertainty margins (min max range for each decade),
- modify breakdown of CAPEX in equipment, installation and civil work, project management, authorizations and right of way
- modify the dependence of each of the five components to the end-indices (e.g. respective contribution of OIL and LAB to installation)
- assume different time evolutions for each of the end-indice: one could easily modify time evolutions of a given commodity (e.g. Aluminum or Copper) or Oil price.
This article is connected to the following related complementary knowledge articles:
Deliverable D3.1 eHighway2050
The possible open access release of the database is currently under discussion.
1 One could mention the FP7 EC-funded IRENE40 project aiming at building a technology database of power transmission systems with simple cost evolution models until 2050.
2 The “learning curve” approach describes how marginal labor costs decline with cumulative production. The “experience curve” generalizes the labor productivity learning curve by including all costs necessary to research, develop, produce and market a given product. The general form of the experience curve is a power curve defined with a progress ratio PR=2-b, where b is the learning coefficient. Thus, for each doubling of cumulative production, the marginal cost decreases by (1-PR). For example with a PR of 90%, doubling of cumulative production within 20 years implies a 10% reduction in marginal cost. It should be noted that the “classical experience curve” includes “all costs necessary to…”: in our study we have separated two effects (the industrial product and the raw material due to its importance for transmission equipment).