A Look at Non-Renewable Generation on Grids with A Lot of Renewable Resources
By Bonnie Marini, PhD
The use of renewable power generation continues to grow around the globe. The challenge introduced with renewable generation is that it can be interrupted by timing and weather, and this variance affects the stability of the power produced. The sum of all generation must meet the demand at the very instant the demand is manifested-simply put, most of the electricity around the globe is produced and used in the very same moment. Without a solution for large-scale, cost-effective energy storage, the only way to fill the gaps in fluctuating generation from renewable resources is by partnering dispatchable power generation.
The California duck curve has become synonymous in the industry with the shape of renewable generation. The renewable duck generation rides on top of the rest of the generation portfolio, which is referred to as the duck pond. This paper discusses the changes in demand for this pond of dispatchable generating resources as they react to the presence and growth of the duck.
Renewable Dispatch and the California Duck
Figure 1 is a picture of a graph from the CAISO that has become commonly used when discussing renewable generation. The picture shows a projection of required non-renewables generation over a 24 hour day. The top of the duck is 2013. The bottom of the duck’s belly is 2020. What can be seen from the duck is that the solar power generated during the day will result in a need for non-renewable resources to ramp down in the morning and ramp up in the evening.
It should be noted that the California duck curve of Figure 1 is a picture of one particular day. Since weather is a big factor, the curve changes every day.
A glance at any of these curves can result in an impression that most of the other generation on the grid will need to ramp on and off in order to support the renewables however in Figure 1 the y-axis is cut off at 11 Gigawatts. It is interesting to note that the duck’s belly is marked “potential over-generation risk”. Clearly this “overgeneration” is not due to the renewables producing more power than the grid demands. It is due to the expected behavior of the resources providing that 12 to 15 GW – the generators that fill the duck pond. Figure 2 is an expansion of Figure 1, showing the amount of generation under the duck.
What changed and what didn’t
Historically, different types of power plants were used for different types of dispatch. Figure 2 shows how the split between base load and fluctuating load plants is changing. Fewer base load plants and more fluctuating load plants are needed to integrate with renewables. The concern is that there are too many base load plants and not enough fluctuating load plants. If the base load plants can’t be turned down or shut down, there is overgeneration. Using the same example graph, moving from the 2013 scenario, the steady base load was about 18 GW. In the 2020 scenario the steady load is about 12 GW, which means that 6 GW of generators had to switch from being base loaded to fluctuating load.
Looking at the curve, we can see that it is not just the amount of fluctuating generation that changes, it is also the amount of time these resources are dispatched. In the 2013 scenario, most of this fluctuating generation is dispatched less than 20 percent of the time. In the 2020 scenario, many of the fluctuating resources are running more than 60 percent of the time. This changes both the economic and environmental impact of these fluctuating resources.
One other factor that has been a point of discussion with renewable integration is ramp rate. Figure 3 shows another example of total demand and non-renewable demand over the course of a day. The blue line depicts total demand and the lower red line depicts the net demand which in this case is defined as the non-renewable demand. This shows that even without renewables, power had to ramp up and down in order to meet demand. The difference between the past and the future is not the existence or rate of the ramp but simply the amount of energy that needs to ramp.
The charts illustrate that increasing renewables
- will decrease the base load non-renewable generation
- will increase the amount of fluctuating non-renewable generation
- will increase the amount of dispatch of the fluctuating non-renewable generation
- will require a similar ramp rate to the fluctuating generation used in the past
The first reaction to deal with this non-traditional problem was using a tradition solution. Some suggested switching these plants to the types of plants that had been used for fluctuating load in the past, for example simple cycle gas fired power plants, but the change in dispatch make this solution a less suitable fit for today’s challenges.
The Right Solution for Fluctuating Generation
The first requirement for a resource to be used for fluctuating generation is that it can support fluctuating operation. For illustrative purposes, as shown in Figure 4, the duck pond can be broken into three different kind of operation.
- Base load = not changing load frequently
- Mid Load = dispatch for a large portion of the day, but can reduce load or shut down for portions on the day
- Peaking Load = dispatch for short operational windows and operate <20% of the time
To support fluctuating load, there are several types of plants that can be used
- Simple Cycle – Fast start, high flexibility, low CapEx, low efficiency, High LCOE
- Conventional Combined Cycle – Slow Start, Capable for Fast Load Changes, Wide operating range, Higher CapEx, High Efficiency, Low LCOE
- Flex-Plant Combined Cycle – Fast start, Capable for Fast Load Changes, Wide operating range, Higher CapEx, High Efficiency, Low LCOE
In addition to technical capability, the decision between technologies depends on economics. For a low dispatch plant, low CapEx is more critical than low LCOE.
For a high dispatch plant, the opposite is true. Low LCOE is the more critical factor. Typically for plants that dispatch less than 10 percent to 20 percent the economics favor simple cycles. For plants which dispatch more, the fuel savings benefit of the combined cycles results in better economics for the high efficiency combined cycles.
Prior to the growth of fluctuating renewables much of the fluctuating load was low dispatch. The solution for this demand was simple cycle. Today’s medium dispatch demands are different.
To gain a broader understanding of how future dispatchable resources will need to behave in order to accommodate increased renewable generation; data developed in two recent studies of future dispatch behavior were evaluated with a specific focus on what types of plants will be needed to accommodate increased renewables.
One study was conducted by the CAISO1 and the other was conducted by the Ventyx Corporation.
The results of both studies indicate that the majority of demand fluctuations will be supported by combined cycles.
In a future grid, with an increase of highly fluctuating renewables, simple cycles will still support low dispatch, peaking demands. Whereas, combined cycles are a better choice for the rest of the pond if they can meet the fluctuating demand – and analysis and history shows that they can.
Conventional Combined Cycles for Fluctuating Demands
Changing load is not only due to changing dispatch of renewable generation. It is also due to constant changes in demand which happen all of the time. Conventional combined cycle technology has been used to meet these changing loads for many years. As shown in Figure 3, the ramps seen with renewables are not expected to be faster than the ramps previously seen in the market – they are only larger and longer.
Figure 5 is a simulation of a winter day on the Huntington Beach grid in California. Many of the plants modeled are not advanced Flex-Plant combined cycles, but are conventional cycles. Figure 6 focuses on the energy provided by combined cycles and shows they are providing majority of the ramping support. The power from combined cycles ramps up and down to cover two peaks during the day. The magnitude of the energy supplied makes it practical to use large combined cycle plants to support this need. The red line on Figure 5 represents the simple cycles. In this case they are dispatched; however they are not used to cover the changes in demand. Their dispatch is rather flat, and the amount of energy dispatched is minimal. It is less costly and more environmentally friendly to use the combined cycles to cover large demand changes so they are used first.
Figure 7 and Figure 8 show a projected summer day, in Huntington Beach. Again, the dispatch of the simple and combined cycle’s show that the larger share of demand change is supported by combined cycles. In this case, the overall demand is high, and the simple cycles are dispatched to meet the peak in demand. This dispatch order on a high renewable grid is similar to the dispatch order on a conventional grid. Combined cycles dispatch first because they offer a lower cost of generation followed by simple cycles meet peaks in demand. There is no indication of a need for more simple cycles to support load changes.
Similar data was extracted for a node in Texas, which has the largest supply of wind power in the US. Figure 9 illustrates that the same phenomenon can be observed there as well. The vast majority of load changes are supported by combined cycles first. Simple cycles are used primarily for peak demand and are not critical for supporting the large ramps in load that were seen in the past, or the even larger ramps in load that are expected in the future. Combined cycles are able to change load quickly and ultimately dispatch first due to the lower cost of generation.
Flexible Combined Cycles for Renewable Support
While conventional combined cycles offer advantages over simple cycles for renewable integration, modern Flex-Plant combined cycles offer significantly more capability. While conventional combined cycles have high ramping capability, they are not designed to start fast and frequently. Newer flexible combined cycle plants can start as fast as a simple cycle while still maintaining full equipment life, making multiple restarts viable even for large combined cycles.
A good example of a flexible combined cycle which uses the advantage of fast start is the Siemens H-Class power plant which has been operating since 2012 in Irsching, Germany. Figure 10 shows an example of the plant operation as it starts quickly in the morning, follows demand during the day, shuts down in the evening, and repeats this pattern the next day. With Siemens Flex-Plants there is no need to trade efficiency for flexibility. This plant exceeds 60% net combined cycle efficiency and can add 500 MW of generation to the grid in 30 minutes. Operating Flex-Plants in the U.S. include Lodi Energy Center in Lodi, California and the Temple and Sherman plants in Texas.
Unlike most simple cycles, combined cycles often have a very large load range, enabling them to ramp up and down without having to shut down and restart.
Today efficiency is not just an economic factor, it is also an environmental consideration. Greenhouse gas production is directly related to efficiency, and recently issued EPA rules are one example of the requirements for future generation. In the EPA’s Carbon Pollution Standards issued this year, in order to manage carbon dioxide production, low efficiency plants are limited in dispatch based on their efficiency. For plants with less than 50% efficiency based on LHV, the plant’s dispatch is limited to its equivalent LHV based efficiency. For example, a simple cycle plant with 35 percent net efficiency may not dispatch more than 35 percent of the time. Again looking at Figure 4, it is expected that a large group of fluctuating plants will be needed to dispatch at higher levels, adding another driver for the use of combined cycles for renewable integration.
Fitting the Plant to the Need
It seems rare when a choice is better in functionality, cost, and environmental footprint, but for high dispatch plants, combined cycles win in all three areas. Flexible combined cycle power plants support the renewable concept by being more efficient, cleaner for the environment and flexible to meet the change in demand.
For low dispatch plants, a traditional approach still makes sense. The lowest CapEx solution with high flexibility is a good fit, and this is typically a simple cycle frame unit.
The growing portfolio of renewable resources has resulted in the addition of a large demand for high dispatch, fluctuating generation. This demand regime is not served by plants which cannot change load, and is not served well by low efficiency plants designed for peaking. This growing market segment needs a solution with high operating efficiency, low LCOE, high ramping flexibility and fast start. The new generation of Flex-Plant combined cycles offer these benefits. Use of the right solutions for these various load regimes will support the environmental benefit of renewable generation while managing the cost of generation with cost effective, high efficiency plants.
Bonnie Marini is director of Product Line Management at Siemens Energy.