 |
The decarbonization of our energy usage might be one of the most important engineering challenge of our times.
Over the last few years, I have been particularly interested in what it would take to build an electric power system primarily on the basis of wind and solar generation and how much it would cost. Based on reading some papers on the topic and doing some back of the envelope calculations, these is a summary of some observations and conclusions I have come to:
- The most relevant metric to evaluate the cost efficiency electricity production is the weighted production cost of a combination of technologies which can satisfy a particular load profile at all times. We could call this average system unit production cost or as it has recently been coined in this paper: LCOLC (Levelized Cost of Load Coverage). While the traditional unit production costs (LCOE) can be simply added up, LCOLC requires the use of an LP-solver.
- For my own calculations, I have mostly assumed a cost optimized combination of 4 technologies: wind & solar generation backed by short and long duration storage (e.g. Li-ion battery storage and hydrogen storage respectively).
- Using current load profiles and solar/wind generation profiles for a variety of European electricity markets (bidding zones) we find a relatively consistent optimal production mix of about 70%/30% wind / solar-PV production. With some minors shifts, this holds true for what seems very different climate zones from the north sea cost (Netherlands or Denmark), Mediterranean (Southern Italy or Greece), Alpine (Austria or Switzerland) or geographically large areas, spanning different climate zones (Germany, France, Spain) - see here or here.
- The optimized configuration requires about 20-30% of energy production from storage leading to curtailment and storage losses of about 20% (see here or here).
- While hydrogen storage is very expensive and inefficient, removing it from the system would increase average costs significantly, mostly by increasing overproduction and curtailment by up to 60-70% (see here).
- Short-term demand flexibility in the order of hours to days e.g. from charging battery electric vehicles of building heat can reduce of remove the value of short-term storage from the system (see here).
- Even if we had a large pool of already paid for battery capacity from a fleet of battery-electric vehicles with bi-directional charging capabilities, this would make hardly a dent in the need for long-duration storage (see here).
- Long-distance interconnectors which might span across climate zones or at least weather systems can help to reduce the need for storage, but again more on the short-term rather than long-term or seasonal storage (see here).
- Nuclear power, which is another candidate for low-carbon electricity generation offers little synergy in combination with wind & solar generation. Nuclear (assumed at 100% capacity factor, constant production) requires its own source of flexible/dispatchable generation for matching the actual load (see here). Below a production cost (LCOE) of ca. 90 Euro / MWh, nuclear could displace the combination of wind & solar (assumed production costs of 65 / 55 Euro per MWh respectively) from the optimal solution (see here).
- Hydrogen energy storage, which is said to be only for the foolish or rich and desperate would seem to play a key role in a future wind & solar based electricity production system - at least for the less than favorable conditions in Europe. Today, hydrogen energy storage is far from being cost competitive with natural gas in the role of long-term flexibility provider of last resort (see here).
- I am deliberately using today's numbers for demand and cost assumptions to minimize the need for further speculation about the future. However it is likely and expected that some technologies would become cheaper over time - specially those which are based on standardized, mass-produced modular components with a current learning rate that can be extrapolated. Current electricity demand will likely increase with the progress of electrification, will likely become more flexible in time (hours to days) by also more seasonal (increasing demand from heating and cooling).
- Static optimization is not sufficient for a robust and resilient energy supply. Additional reserved and redundancies are needed in the system to handle shocks and crisis situations. This would likely increase the need for additional long-duration storage with a time-horizon across multiple years. This recent paper shows an interesting approach to jointly optimize for both concerns in a single model.
- In these discussions we are ignoring any concerns about transmission and distribution costs. The zonal design of the European electricity markets, transmission is treated as a separate concern out side the market and often treated as a natural monopoly. Transmission and distributions cost are dominated by the last mile of the distribution network, which depend on settlement structures and population density.
- Most energy infrastructure is heavily subsidized. The total cost of production outlined here is typically covered through a combination of energy market revenue and a variety of direct and indirect subsidies.