Is Renewable Energy Ready to Run the Grid? Miracle(s) needed

Levelized cost of Energy (LCOE) is the benchmark on which we celebrated recently that renewables (wind and solar) are now cheaper than all other forms of energy production. This however comes with a significant caveat: as opposed to other types of energy, renewables are intermittent and to not supply energy per demand. Photovoltaic Solar (PV) for example, produces power on average 10-30% of the time. So although this power is cheap, it is not being produced 90-70% of the time. But demand is on, all the time. Extending renewable power with storage for even a few additional hours a day makes it more expensive than non renewable sources of energy.

Figure 1: PV power production (in June and January) and average grid demand per hour.

Demand varies hourly, weekly and monthly but daily fluctuations are about +/-20%. Source: author's PV system and Israel electrical company data (2006).

The way to solve the discrepancy between limited production time and always on demand is to add energy storage, thus bringing energy from noon when it is abundant (for solar) to other parts of the day when it is needed (basically 24 hours a day- for simplicity let's assume power demand is constant). To make things even simpler, we will ignore the randomness of seasons and weather events creating weeks with very little production, and of partially cloudy skies interrupting noon solar production. We will assume a fix production each day- several hours around noon. For sunny countries, solar produces power about 20% of the time. This translates to 4.8 hours each day on average. There are many storage technologies, but let's focus on the only one currently being installed in wide scale commercial projects: the Li Ion battery. As an example, let's double the hours a day we can send solar power to the grid. We add a battery that is large enough to store the entire energy created by our PV system per day. Alas, now we need solar production to double: half goes to supply the demand as before, and the other half is used to charge the battery during peak production hours. After production hours are over, the battery is discharged to supply the grid. So in general, to supply 24 hours of energy, with solar producing only 20% of the time, we need to increase PV production by a factor of 5x, and have batteries that store 80% of the grid energy consumption per day (!). To put this into perspective: a very small affluent country would need ~300GWh a day of energy. The biggest battery installed so far is about 0.6GWh (600MWh, California). Let's put aside the sheer magnitude of this colossal storage system, with its material demands (probably we do not have enough Lithium in the world, with EV competing for it), land requirements and impact on the environment (extraction is not as clean as we would like it to be). Let's focus only on the financial costs and assume the following: Note that Li Ion storage LCOE is currently higher, I have used the 2030 best case projection from DOE estimates. So, when and how is it cost effective to add storage? how does a PV+storage only grid look like?

Figure 2: simplified cost calculation for increasing power output hours in a day.

PV reaches is quoted LCOE when used "naturally", without storage. Increasing solar production beyond that increases costconsiderably. We see the under these assumptions, a 24 hour PV + Li Ion storage costs above 170$/MWh, ~1.5x more than nuclear energy and ~2.5x more than coal. It is clear why developing countries are doubling down on coal power plants, the most carbon intensive form of energy production. This is not to our climate's advantage. So, with this simplified model, under what assumption can we have a renewable grid be more cost effective than nuclear power? We need the cost of storage to be 20$/MWh, 10x less than most optimistic DOE estimations. But even if this miracles will happen in time for 2030, we are still far from a solution: the calculation assumed a no randomness induced by weather. If we have a week of clouds and low wind, our grid dries up after 1 day. We also need another solution for transferring energy from summer to winter- Li Ion batteries will not work for that. And we still need to build backup power plants that increase carbon emissions and costs. The numbers seem to indicate that for 2030, renewables can not decarbonize electrical production- not nearly close (we are 8 years away, even if tomorrow we find a miracle storage solution- it will take years to develop and implement globally). As supporting evidence, let's compare Germany and France: While Germany has invested ~200 Billion Euro over the last 20 years in renewables it still has almost twice the energy production carbon footprint compared to France (spoiler: 70% of French electricity is from Nuclear power plants and they have announced 6 more nuclear power plants to be build, while Germany is committed to closing its last 3 plants by end or 2022).

Figure 3: Carbon intensity in energy production. Source : Our world in Data

In summary:

• Without storage technology breakthroughs, renewables are not a solution for a ~constant load grid requirements.

• We need to develop additional alternatives for decarbonizing grid power.