Most individuals operating in the renewable energy sector are aware of the challenge of “intermittency” found in wind, solar, and other renewable sources - i.e. wind/solar electricity output is variable by nature and thus cannot serve as a guaranteed, baseload power source.
Below, I discuss some of the concerns that scientists have raised regarding PV solar intermittency specifically, and then cover new studies indicating the broad potential of wind power to mitigate these concerns.
One issue with distributed photovoltaic solar involves the “waste” of solar PV-generated electricity, which theoretically occurs if PV reaches large US penetration levels. According to two studies from Paul Denholm and Robert Margolis out of NREL (unfortunately only one is available publicly):
The intermittency of solar PV...presents a set of critical challenges with respect to integrating PV on a very large scale into the electricity grid. Ultimately, this intermittency may limit the potential contribution of PV to the electricity sector.Once PV provides 10%-15% of the overall electricity portfolio of a traditionally structured grid, the inflexibility of the current baseload system ensures that any additional PV generation will mostly be wasted.
This is due to the nature of the current U.S. electricity grid system and the difficulty for PV solar to match volatile demand due to its intermittency. Large baseload plants are limited in how far and fast they can drop or boost capacity to match demand (due to various costs, efficiency and timing reasons) According to these studies, past a certain level of penetration (10%-15%) PV solar can only augment existing capacity, and not replace it. On moderate or low demand days, PV electricity would be wasted in order to avoid interfering with baseload plant operations.
The two researchers found that under high penetration levels and existing grid-operation procedures and rules, the [utility] system had excess PV generation during certain periods of the year that increased PV costs—that is, the PV electricity had to be dumped. The limited flexibility of baseload generators, which cannot respond to rapid changes in load, produces more unusable PV generation when PV provides more than approximately 10%-20% of a system’s energy. [source]However, two wonderful posts at The Energy Blog pointed me to new studies that may challenge this conventional wisdom.
The first study, out of Stanford, makes the claim that connecting multiple wind farms with some amount of geographic diversity creates a large and diversified wind power resource with enough reliability to serve as a baseload power source.
Interconnecting wind farms with a transmission grid reduces the power swings caused by wind variability and makes a significant portion of it just as consistent a power source as a coal power plant. This study implies that, if interconnected wind is used on a large scale, a third or more of its energy can be used for reliable electric power….I’m reminded of both diversification theories in finance, and aspects of the Central Limit Theorem, in looking at these explanation. Is there a statistical theory involving variance in systems, and being able to mitigate that variance with diversification of inputs? If so, could one of my much smarter readers please fill me in?
…The researchers used hourly wind data, collected and quality-controlled by the National Weather Service, for the entire year of 2000 from the 19 sites. They found that an average of 33 percent and a maximum of 47 percent of yearly-averaged wind power from interconnected farms can be used as reliable baseload electric power. These percentages would hold true for any array of 10 or more wind farms, provided it met the minimum wind speed and turbine height criteria used in the study.
The Stanford study also discusses the concept of interconnecting wind farms to a common point, which improves efficiency and cost. Essentially, this strategy recreates the centralized power generation and distribution model which comprises much of the current U.S. electricity system. Those seeking to rebuild American electricity grid infrastructure around distributed power models may be disappointed. But power generators, utilities, and the various government entities that oversee/collaborate with these groups, may be more comfortable with centralized energy concepts. Equating a portfolio of interconnected wind farms to a large 1GW coal plant (in terms of T&D issues, etc.) could improve understanding and acceptance in these circles.
However, I was even more fascinated by a recent study from the Cambridge-MIT Institute (again h/t The Energy Blog), which focused on the potential outcomes for strategic energy security from developing a diversified electricity production system. This highly readable study posited some insightful results:
- in order to explore the potential for geographic diversity, the study reviewed the correlation between wind speed and distance, and found that “sites far apart exhibit very low cross-correlation”. At 600km distance, correlation (r) was about 0.30, while at 800km, it dropped to 0.20.
- in exploring “the percentage of UK sites that have simultaneously experienced calm conditions for one hour” from 1982-2000, it found that “there has not been a single hour in the last 15-20 years when conditions of total calm were experienced right across the UK”. Meanwhile, calm conditions lasting one day, “affect less than 2% of the UK with the remaining 98% of the UK experiencing wind at these times.”
By taking a planned approach to the development of wind power, the impact of distance on correlated output can be fully exploited within the UK, improving the reliability of wind power and minimising the additional backup capacity required due to the presence of wind power on the network. The additional backup required to support 20% electricity generation from wind is estimated at around 4GW.However, perhaps most interesting to me was the attempt by the study to model the impact of various climate policy scenarios on electricity diversity.
Where no emissions target is imposed, there is a decline in diversity in all three scenarios. This decline is driven by an increase in the proportion of generation accounted for by natural gas. The implication of this fall in diversity is an increase in insecurity, as the electricity system becomes more exposed to one fuel source. By contrast, under an emission target of 60% there is a substantial increase in diversity under all three scenarios as the dominance of natural gas goes into decline.Ensuring smart and strategic policy, and incentivizing markets appropriately can aid in the pursuit of energy independence and large levels of penetration for renewable energy.
These basic results prompt two observations. First, low carbon scenarios appear to be associated with higher diversity. Second, these results are largely driven by changes in the share of generation accounted for by gas.
Denholm and Margolis’ solutions for the PV solar penetration challenge include increasing the “flexibility” of the conventional system, meaning that base load plants can cycle down to lower levels, “dispatching” load more efficiently with smarter appliances or developing more effective energy storage system.
The Cambridge/MIT study suggests a “concentration charge” which could either “levy a surcharge on [energy supply companies] in proportion to the diversity index of their overall portfolio” or levy the charge “source by source to reflect the concentration of each source to the system.
So is the intermittency issue no longer a problem if you diversify your renewable energy “portfolio” across a geographically broad enough environment? Great news if so...