The energy sector today is responsible for approximately one third of all global greenhouse gas emissions. According the International Energy Agency, the share of electricity in total final energy consumption has grown rapidly since 2000, from just over 15 % to 19 % in 2017. In the combat against climate change, great hopes therefore rest on zero-carbon, renewable energy sources like solar and wind to decarbonize electricity systems around the world. In order to realize this ambition however, their share in the generation mix must increase substantially, which for some time now has been raising questions about the inherent variability of these resources.
The physical properties of the electric grid dictate that supply and demand of electricity in the system must be in near-perfect sync at all times. A mismatch between generation and consumption of electricity beyond a very small tolerance level, even if just for a few moments, can lead to a breakdown of the entire system. Therefore, certain balancing measures have always been necessary to keep the grid stable, since demand can never be forecasted perfectly and production assets can always malfunction unexpectedly.
On the one hand, generation assets need to be committed ahead of time for dispatch in a way that ensures scheduled consumption loads can be met (i.e. that enough supply is provided to satisfy forecasted demand at all times). On the other hand, there must be physical measures to balance the grid in cases of unexpected short-term deviations, like an unusually hot summer day or equipment failure at a power plant. Historically, power plants themselves have been providing these physical balancing services known as flexibility, mostly by ramping their production up or down at the bidding of the grid operator.
Energy storage is another potential physical balancing measure. It has received increasing attention in recent years as the share of variable generation assets has risen and the cost of battery storage has dropped sharply. According to Lazard’s Levelized Cost of Storage (LCOS) Analysis, the range of unsubsidized LCOS for Lithium-Ion batteries across a range of use cases has shrunk from 211-1,596 $/MWh in 2015 to 108-1,152 $ per Megawatt hour in 2018. Yet while storage will be an increasingly important enabler for more variable renewable generation, it is not the only one available.
A recent study done by E3 and sponsored by First Solar analyzed Tampa Electric Company’s (TECO) portfolio under different solar PV penetration scenarios. It found that under the currently dominant “must-take” approach to grid integration of renewables, where the grid operator must allow them to feed into the grid whenever they produce, TECO’s system can physically only absorb up to 14 % of solar PV before over generation would threaten to destabilize it. The study shows that other generation assets in TECO’s portfolio, mostly thermal power plants, can only provide a limited amount of flexibility to balance the increasing variability in the system as more and more solar is added.
By simulating solar penetration levels between 0 and 28 %, the study finds that commercial dispatching tools alone could allow for at least 28 % of solar PV in TECO’s portfolio. While acknowledging that energy storage could also achieve this, it focuses on the role of scheduled flexibility for utility scale solar PV plants by investigating several different scenarios. In the “must-take” scenario, solar is added only to the point where over generation becomes an issue, effectively capping it at 14 % in TECO’s case. In the “curtailable” scenario, solar will be curtailed in cases of over generation, which means more of it can be added without destabilizing the system, but also that its marginal value diminishes sharply since curtailed solar will produce less power per dollar invested.
This is where the study finds, somewhat counterintuitively, that planning for potential solar curtailment ahead of time reduces the actual need for curtailment. In this “downward dispatch” scenario, the grid operator factors the potential for solar curtailment to provide downward flexibility into its ex ante dispatching decisions. Since thermal generation units can only operate flexibly within a certain range, more of them have to be dispatched overall in order to provide more flexibility. Thus when solar can provide some of that flexibility, less thermal generation needs to be committed, which in turn frees up more room for solar production. As result, solar retains its economic value at high levels of solar penetration in the “downward dispatch”, compared to the “curtailable” scenario.
Although the results of the TECO case are not directly transferable to cases with different generation mixes or regulatory environments (TECO operates its electricity system as a Balancing Authority), they hold important and universally relevant lessons. Energy storage, while vital and increasingly cost-effective, does not have to shoulder the burden of more renewable generation alone. Great potential also lies in (re-)organizing electricity systems to make dispatching decisions more effective and efficient in coping with this new reality. Yet, future research will have to investigate the incentive structures in different regulatory settings and under different generation mix scenarios, as well as interactions with technologies like energy storage, in order to identify optimum pathways towards zero-carbon electricity systems.