Updated: Feb 24
In part III of this series we looked at how existing dispatch and contractual arrangements are inadequate for the integration of renewable power much above about 15% penetration. This arises if renewables are regarded as a “must run” source, as then the grid operator has to be able to call upon significant amounts of short-notice thermal capacity to provide back up if the forecast amount of renewable power is not available at the time, or the grid demand is higher than expected. In the past, certainly for many European grids, the task of balancing has been made easier by the historic “over investment” in thermal power capacity, to provide adequate reserves on the grid, in circumstances when the costs of maintaining this reserve could easily be passed on to the consumer. Recently however, such spare capacity has become scarcer, for example with the retirement of coal-fired plant and the reluctance on the part of investors to construct new gas plant in the uncertain contractual conditions of a rapidly changing market.
In this final part of the series, we look at the efforts being made to work around these restrictions to allow renewables to gain greater grid penetration whilst still allowing full operation of the grid.
The first step which many providers such as Germany, Denmark, California and Hawaii have taken is to renegotiate the Power Purchase Agreements (PPA) between the grid operator and renewable generators to allow the curtailment of renewable power, rather than to purchase it on a guaranteed “must run” basis. This change goes a long way to avoiding oversupply of power into the grid, but requires contractual creativity to ensure the continued viability of existing plant and the attractiveness of new renewables investment.
Coupling these contractual changes with:
improvements in the forecasting of wind and solar output, both from “day ahead” and several hours ahead predictions;
averaging output from an increasing number of renewable installations; and also
maintaining or improving grid connections with neighbouring countries or regions to trade surplus or shortfalls in power;
gives the grid operator greater confidence to use more power from renewable sources. In the case of Denmark, for example, these curtailment provisions have allowed renewables to increase their share of the annual power market to around 42%; a considerable advance.
The economic consequences for the overall supply portfolio and the continued viability of both renewable generation and any backup thermal generation still required are complex, and the subject of current studies such as the one undertaken by First Solar and Tampa Electric Company in the US.
The next logical step after making renewables “curtailable” is to make large scale wind and solar plants “dispatchable”; that is for the grid operator to have sufficient confidence in a particular level of renewables output in a given period that it can plan on dispatching that level as part of the overall power mix. In practice, this means that the level of renewables dispatch will be significantly below its forecast level of output, to make allowance for “real time” uncertainties in both supply and demand.
An example of this reduced dispatch amount, for a solar farm, is shown below:
By planning to dispatch the solar plant (red line) at a level between its likely maximum production for the day (dark blue line) and minimum output (light blue line), the grid operator thus provides both “footroom” should demand drop unexpectedly or “headroom” should extra supply be required. The latter requires that the system operator has sufficient confidence that additional solar production is available when called upon at short notice, of course. This arrangement will therefore lead to a natural conservatism in dispatch.
In the case of solar, generation plant can ramp up or down far quicker than conventional thermal plant, meaning that using this approach could actually improve the flexibility of the system as a whole.
The current economic studies of this sort of plant dispatchability indicate that it can add value to the grid up to around 28% (solar) penetration, thereby reducing the requirement for thermal power as a backup and hence possibly deferring the need for future investments in thermal generation capacity.
A schematic of the non-dispatchable, versus dispatchable, use of solar in this way is shown in the figure below. Not only is solar curtailment reduced in this example (dark red area) but also the need for gas fired power is reduced in the significant periods of the day, as shown by the height of the two arrows on the left and right diagrams.
The same logic applies broadly to wind power on the system, although in a more complex way, as allowance has to be made for local uncertainty in wind generation across the system.
The dispatch of renewables in the manner described above is the next possible step in the evolution of the grid system.
A step further will be the addition of “grid scale” storage to allow even greater flexibility. This storage has hitherto been in the form of pumped hydro, where such capacity exists. Pumping water up and down between two reservoirs has been the traditional means of “load displacement” between to accommodate the difference in time between demand and supply peaks. Pumped hydro can additionally be used to accommodate the uncertainties of renewables output, as has been touched upon in previous parts of this brief.
Storage could also be in the form of batteries and/or underground compressed air storage. The former is, with current technology, difficult to scale up, whereas the latter requires a suitable salt dome which can be hollowed out and sealed.
This will be the subject of future "insight" briefs.
Nelson, J. et al, “Investigating the Economic Value of Dispatching Solar Power Plants”, September 2018
Solar Power Europe, “Grid Intelligent Solar: Unleashing the Full Potential of Utility-Scale Solar Generation”, 2018