Electric grid flexibility and costs of integrating wind and solar
This week at the American Society of Mechanical Engineers 5th Annual Engineering Sustainability conference, as usual, there were many good presentations on various topics from concentrating solar power (CSP) to assessing wind farm design with multiple turbine sizes. One interesting presentation and paper was from Paul Denholm of the National Renewable Energy Laboratory. Paul’s paper is entitled “Enabling technologies for high penetration of wind and solar energy.” Here, high penetration means (wind and solar photovoltaics (PV) and concentrating solar power with and without thermal storage systems).
As indicated by Paul, the limit of system flexibility in the grid will limit integration of renewable generation before pure capacity constraints will occur. This is because the renewable generation will begin to get curtailed before load mathematically exceeds the capacity on the grid due to the fact that the combination of dispatchable sources (hydropower, nuclear, coal, and the various natural gas prime movers) will run into difficult economic choices regarding staying on at minimum levels. That is to say, thermal generators, primarily nuclear and coal, are not inclined to turn completely off at some point during the day and then come back online during another part of the same day (or 24-hr period). There are physical reasons for this as making mechanical parts get hot and then cold too quickly can cause premature failures. This is not to say that engineers cannot design coal plants to ramp up and down more quickly with less wear and tear, but coal plants on the ground now are stuck with much of their design characteristics. And while coal plant operators indicate in person that those plants can ramp up and down very quickly (10s of MW per minute), the data on how much they can really do this on a continuing day-to-day basis is not available. It is also not clear how much industry even knows the answer to how much cycling a coal plant can handle. Nonetheless, Paul Denholm’s calculations from his conference paper being to give us some insight into how existing grid flexibility has an influence on marginal grid prices as more renewables are integrated into the grid. Here he defines “flexibility” as the conventional thermal fleet’s ability to ramp and rapidly follow load. For the most part grids with high capacities of hydropower and natural gas are more flexible than those with high capacity percentages of coal and nuclear. Paul discussed results of his running a unit-commitment dispatch model , a model that determines the most economic manner in which to run an electric grid fleet while considering operating costs as well as costs associated with turning on and off the generators. He compared the increased levelized cost of the total mix of renewable generation as a function of both the flexibility of the grid and the level of renewable integration. For example, if 50% of the grid energy is from wind and solar (PV and CSP) and the rest of the grid is “100% flexible”, then the effective levelized cost of the renewables is only about 4% higher than if there were not costs of integration. However, at the same 50% integration level with an 80% flexible grid, the effective levlized cost of the renewables was found to be 50% higher. The increased cost comes from curtailment of renewable generation that occurs because thermal generators will choose to stay at a minimum level of operation even when uneconomic for them to do so at that specific time interval. That is to say the costs of turning off and turning back on are higher than the costs of operating below cost for a few hours.
Because much of the electric grid is “inflexible”, this is an important area to consider for continued integration of renewables. Thus, this is one of the main reasons for consideration of storage technologies to reduce curtailment of wind and solar generation: renewables will be curtailed even before the total electrical demand becomes less than instantaneous renewable generation due to the inflexibility of the thermal power plants.
In the Electric Reliability Council of Texas, there is a relatively large quantity of natural gas generation with combined cycle systems dominating. Approximately 65% of ERCOT capacity is in natural gas-fired power plants with nearly 8% of total electricity now provided by wind power. While it is foreseeable that ERCOT could have 20% of total annual generation from wind power within 10 years (approximately a doubling of the current 9.5 GW of capacity is needed along with completion of existing plans for transmission lines), 30% total generation from wind and solar technologies seems further off without some new social or political drivers. However, ERCOT has already seen many hours of operation over the last couple of years during which wind generation accounted for over 20% of generation on the grid (mostly during times when low demand in spring and fall coincide with high seasonal winds during early morning hours). This is a time when the grid is technically less flexible (e.g. when the grid is more dominated by coal and nuclear generation than natural gas) but in which the spare natural gas capacity exists to come online if needed. Thus, because ERCOT has the luxury of a large capacity of natural gas combined cycle plants to meet summer peak loads, it is relatively easy to handle large wind penetrations during times of low loads. If there was a large quantity of wind power during times of high demand, then it could present new problems.
Denholm’s analysis in his ASME paper was somewhat simplified in its procedure (e.g. neglected transmission constraints) to more readily explore the issues regarding high percentages of wind and solar integration) also indicates that we would likely not see much increase in the effective levelized cost of renewables on the grid at penetrations below 20% total generation. We do not have examples of over 20% wind and solar on a single grid or within a single ISO in North America, and ERCOT has not yet seen significant costs associated with grid inflexibility. However, ERCOT has seen significant increased in effective levelized cost of electricity from wind power due to curtailments from lack of transmission capacity. Given that the Public Utility Commission of Texas has already set in motion the building of new transmission lines to access wind power regions of Texas, it is anticipated that the wind-grid congestion issue will be resolved in a few years.
While Denholm’s analysis is quite forward-looking into the future, it is still important to consider what the world of high penetration of renewables will really look like as much as we can understand with use of existing technologies and operational principles. This will help us anticipate concerns from both renewable and thermal (coal, NG, nuclear) power plant operators to properly distribute investment costs to consumers in as equitable and inexpensive manners as possible.
TrackBack URL for this entry: