By Brian Elwell and Kieran McInerney
In today’s dynamic energy industry, the need for flexible, efficient electricity generation is increasingly important. The days of predictable peak demand patterns are behind us. As wind and solar energy sources achieve greater market penetration, their intermittent energy supplies present challenges for Independent System Operators (ISOs) to balance loads and maintain frequencies. These challenges are often intensified at sub-transmission voltages.
As demonstrated by recent market trends, reciprocating engine generators are well suited for flexible peaking and intermediate generation needs in the 20-300 MW output range. Reciprocating engines offer competitive heat rates and multi-shaft reliability to compete in energy markets, plus industry leading ramp rates and startup times to compete in ancillary services markets. The information presented herein highlights market factors driving the growth of the reciprocating engine market and compares reciprocating engines to other generating options.
|The Fairmont Energy Station is a 25-MW project equipped with four Cat G16CM34 generator sets. The plant was commissioned in 2014. Photo courtesy: Caterpillar Inc.|
Utility-scale engine models include the 9-10 MW and 18-20 MW unit classes. These are heavy duty, medium speed (514 – 900 rpm) engines that can easily adapt to grid-load variations. Table 1 shows the expected net output and heat rate values from three similarly sized peaking plant options: a 220 MW reciprocating engine plant based on the 18-20 MW unit class, a 2x 100 MW aeroderivative gas turbine (GT) plant, and a 1x F-class frame GT plant. A 1×1 F-class combined cycle is also included to demonstrate how these options may compare to an indicative option for intermediate load dispatch.
The reciprocating engine plant has the lowest full-load heat rate among the peaking options, and this difference is even more pronounced at part load operation. There are two ways for the engine plant to achieve 50-percent plant load. First, half the plant can be operated at full load, which will essentially maintain the full load heat rate, depending on the auxiliary loads still running. Second, if all reciprocating engines are ramped down to 50-percent load simultaneously, the resultant net heat rate is still better than the frame GT full-load heat rate. This trend is illustrated in Figure 1, which shows the net heat rate curves at summer conditions.
|Summer Heat Rate Curves|
|Based on summer conditions of 95°F and 50 percent relative humidity.|
Reciprocating engines can start up and ramp load more quickly than most gas turbines, but it should be noted that the engine jacket temperature must be kept warm to accommodate start times under 10 minutes. In addition, reciprocating engines are generally more tolerant of altitude and ambient temperature than gas turbines. With site conditions below 3,000 feet and 95°F, altitude and ambient temperature have minimal impact on the electrical output of reciprocating engines, though the efficiency may be slightly affected.
When operated and maintained according to manufacturer recommendations, modern reciprocating engines commonly exhibit availability factors of 95 percent or better. Since there are often multiple engines at a given site, maintenance outages can be staggered to avoid taking the entire plant offline. Similarly, an unplanned outage event for a single engine will not force the entire plant offline.
The case for flexible generation begins with the changing makeup of the electricity market. In 1999, over half of U.S. energy was produced by coal fired generators, according to data from the Energy Information Administration (EIA). In that same year, wind and solar energy were essentially negligible at approximately 0.1 percent (combined) of total generation. Since then, environmental and economic drivers have continuously changed the composition of U.S. energy supply. In 2015, coal energy fell to 33 percent of total generation while the combined impact of wind and solar rose to 5.3 percent. This may not seem significant when considered nationally, but renewable integration is unevenly distributed in the country, so there are substantial impacts in areas where renewable penetration is high. For example, wind accounted for over 9 percent of the installed capacity and 7 percent of generation in MISO in 2015. In ERCOT, wind accounted for nearly 12 percent of generation in 2015 and set a record for instantaneous penetration of 45 percent of demand in February 2016. However, in PJM, wind accounted for less than 4 percent of generation in 2015.
Meanwhile, the makeup and behavior of energy consumers must also be considered. During the latter half of the 20th century, the electricity consumption trends for industrial, commercial, and residential customers steadily increased in parallel paths, as shown in Figure 2.
|Retail Electricity Sales by Sector|
|EIA 2014 Annual Energy Outlook|
However, it is important to notice that electricity sales to industrial consumers flattened out in the 1990s while commercial and residential trends maintained their upward slopes. Industrial demand is generally understood to be a primary component of baseload electricity demand, while commercial and residential users typically set the peak demand levels. Therefore, a simple review of Figure 2 suggests that baseload demand has remained steady in recent years while peak demand continues to rise. Further, because energy costs often represent a significant annual expense for industrial consumers, they are more likely to have financial incentives to reduce energy consumption. Annual energy expenses for commercial and residential consumers are typically less significant in proportion. Except in areas where energy costs are significantly higher than the national average, there is less financial motivation for those consumers to change their consumption behaviors. This suggests that the trend in Figure 2 is likely to continue, which means the delta between the daily peak demand and the baseload demand may grow.
In the past, increased peak demand was a relatively simple problem to solve, but it gets more complicated when considering the impact of renewable energy integration. Wind and solar sources are rapidly increasing their shares of annual energy production, but the timing of this production can be a challenge for load serving entities (LSE) and ISOs. Because they are intermittent resources, it is often difficult to match the load profile in areas with high renewable penetration. When the wind stops blowing, there is a need for a reliable generator with fast starting and ramping capabilities to quickly fill that energy void.
Wind energy is generally more consistent at night than during the day, and also tends to be stronger during the spring and fall than during the summer. Wind energy sources are less likely to provide maximum benefit during summer days when peak demands are highest.
Solar generation is obviously at its best during summer days, so these sources can help reduce peaks when the sun is shining. However, increased solar penetration creates an even greater need for flexibility among the generating fleet. Figure 3 demonstrates how significant implementation of solar resources can reshape a daily demand curve by displacing typical peaking, intermediate, and baseload resources.
|“Duck Curve” – Impact of Significant Solar Integration|
|Source: NERC and California ISO: 2013 Special Reliability Assessment: Maintaining Reliability While Integrating Variable Energy Resources – CAISO Approach|
The blue line shows a traditional demand curve with a consistent base load, intermediate plateau, and late afternoon peak. The yellow and green lines show the impacts of solar and wind, respectively. Finally, the red line shows the net resultant load that must be met by dispatching the generating fleet. Note that the red line has two daily peaks, which have led many to label this a “duck curve” because the daytime shape resembles the back of a duck. This curve demonstrates the importance of a fleet’s ability to ramp up and ramp down quickly to react to the impact of solar generation.
Localized grid factors are also driving the need for distributed energy resources, and reciprocating engines are competitive in that arena. Electric loads are typically connected at voltages under 230 kV, and extra-high voltage transmission systems often do not provide voltage support to lower levels. Renewable energy sources are also commonly connected under 230 kV, and often at distribution voltages. In areas where renewable sources are clustered and transmission constraints exist, flexible generating assets can help mitigate the impacts of intermittent electricity resources. For example, wind farms are common in the middle of the country, in areas that also tend to be sparsely populated, so the wind farms may be a significant portion of the local generation profile. LSEs need to balance those systems, but they likely have a much higher area control error (ACE) than an ISO, so flexible generation assets may be critical at lower voltages. In those areas, the fast start times and ramp rates of reciprocating engines, along with the ability to scale the installation linearly by reducing the number of engines at a given site, are well suited to “follow the wind,” maintaining desired output and frequency.
The technological benefits of reciprocating engines are evident for flexible peaking applications, but supply and demand drive the success of the plant, as demonstrated in an indicative economic analysis. For the example plant options introduced above, generation revenues were evaluated based on a production cost model (PROMOD IV) simulation using historical data from 2011 – 2014 in MISO, ERCOT, and PJM. Essentially, the evaluation shows generation revenues according to the dispatch results during the study period.
Figure 4 shows the expected annual operating and maintenance (O&M) costs plus major maintenance costs. Fuel is excluded from variable O&M. On a $/kW basis, there is little differentiation among fixed O&M costs for the peaking options.
Variable O&M rates are highest for the aeroderivative option, largely due to costs related to demineralized water consumption. Note that the aeroderivative model can be selected with dry combustors and fin fan cooling to minimize water consumption. That model would be expected to have variable O&M costs of approximately $1.50/MWh, but the output at summer conditions would be more than 10 percent lower. The reciprocating engine plant requires very little water, measured in gallons per week, for cooling loop makeup. However, there is routine, minor maintenance on the engines plus SCR reagent and catalyst replacement considerations.
Figure 5 shows the comparative energy revenues based on the model results. The results from three example ISOs are broken out by color shading. The superior heat rate of the reciprocating engines allows for a higher capacity factor than the other peaking options, and therefore better energy sales. The combined cycle was shown for indicative comparison between peaking and intermediate dispatch applications. There are applications for intermediate dispatch of reciprocating engine plants where a combined cycle may not be financially prudent. In direct competition at higher capacity applications, the combined cycle will generate more revenue due to its superior heat rate.
|Energy Market Snapshot|
Depending on the ISO, capacity markets and ancillary services markets (ASM) offer additional opportunities to generate revenue. Ancillary services commonly include frequency response, spinning reserves, and non-spinning reserves. Frequency response is typically the most lucrative opportunity, and it also requires the most rapid response times. Reciprocating engines are well suited to compete among fossil fuel options for frequency response. Annual revenues from ASM represent a smaller piece of the total revenue pie, but they are a potential tiebreaker when choosing the most favorable generating option for an application.
Capital costs obviously play a major role in an economic analysis because they are the hurdle to profitability. Of the peaking options shown, the F class GT has the lowest expected cost per kW output ($/kW), based on an engineer, procure, construct (EPC) contract methodology. The EPC cost for a generic simple cycle F class plant is approximately $550/kW, while a generic 220 MW reciprocating engine plant is approximately $950/kW. The capital cost difference is analogous to the energy density of the plants themselves. The frame GT generates over 225 MW from a single machine, while the reciprocating engine plant requires 12 units to achieve a similar capacity. In simple terms, the engine plant requires more material to generate the same output, but the multi-shaft design allows for scalability and right-sized plant design in addition to the operational flexibility and superior heat rates outlined above. In any evaluation, it is important to evaluate the costs and benefits for the specific application.
Each opportunity and each ISO is unique. The market for reciprocating engines is rapidly expanding because of flexibility and performance benefits. Full load and part load heat rates, startup times, ramp rates, and multi-shaft reliability all favor the reciprocating engines over gas turbine alternatives. In addition, reciprocating engine designs allow for scalability and right sized solutions, especially for applications in the 20-300 MW range. As the industry accelerates toward an even more dynamic and uncertain future, the market for proven, flexible generation will continue to grow.
Brian Elwell is EPC Project Manager at Burns & McDonnell. Kieran McInerney is a Development Engineer in Burns & McDonnell’s Energy Division