COMBINED CYCLE TRENDS: PAST, PRESENT AND FUTURE

A NEW BENCHMARK — IN LIEU OF ISO BASE LOAD EFFICIENCY — IS NEEDED FOR NATURAL GAS-FIRED COMBINED CYCLE PLANTS

As the saying goes “it is difficult to make predictions, especially about the future.” Nevertheless, it is unavoidable in cases such as investing in stocks or betting on horses. In those instances, of course, one is always cautioned that past performance is not necessarily indicative of future results.

While the caveat is true enough, past performance is the most salient piece of data we have. This is certainly true for gas turbine combined cycle (GTCC) facilities. As such, drawing upon the historical data, this article provides a glimpse into the future of GTCC technology. Using the traditional metric of ISO base load efficiency and fundamental thermodynamic arguments, it will be shown that a limit is about to be reached. This in turn suggests that we do need a more meaningful performance benchmark to evaluate the future of GTCC technology.

Before delving into data, it is instructive to look at predictions made by engineers and scientists on the same subject going back as far as early 1980s (Figure 1). Extrapolation of data from OEMs straddles the technology curve. This is to be expected because, per their coincidence with the U.S. Department of Energy’s (DOE) Advanced Turbine System program target, data points from OEMs A and B reflect most optimistic theoretical efficiencies whereas those from OEM C reflect commercial performances.

On the other hand, a 20+ years-old paper from academia seems to be right on the money with its prediction of today’s state of the art. The most puzzling item in Figure 1 is the severe underestimation of future GTCC performance in a 1983 paper —– even with water-cooled hot gas path components (a prominent feature of the DOE’s High-Temperature Turbine Technology program in the 1970s) that were shown to be capable of achieving 1,600°C turbine inlet temperature (TIT). This is a dramatic illustration of ignoring the contribution of the bottoming cycle to GTCC performance; there is only so much one can achieve just by burning more fuel.

Another instructive exercise is to look at the evolution of fossil fuel-fired (mostly coal) steam-driven power generation technology (Figure 2). The early steam engine and Parsons’ fin de siècle steam turbine data are from W. S. Jevons’ prescient treatise “The Coal Question” and Aurel Stodola’s seminal work “Steam and Gas Turbines.”

The modern age data (containing historical and forecasted numbers) is compiled from various sources such as IEA databases published in several reports and articles. It shows the classical stages of the technology life cycle or the “S curve,” ending in the mature phase capped by about 45% net efficiency. That barrier seems unlikely to be broken for quite a while (maybe never considering the stringent limits imposed on the generators by regulatory agencies).

The technology trend from the 1980s to the present for GTCCs is illustrated in Figure 3. Data from trade publications and other public sources (mostly rating performances but some plant data as well) is reasonably well represented by the following exponential formula (Y is the year):

Based on this formula, one can reasonably expect ~62% by 2025 and ~62.5% by 2035 (ISO base load rating performance). Allowing for a margin of error, these numbers can be bumped up by one percentage point.

But it is important to consider the difference between rating performance and reality. To do that, let us look at the performances of gas fired GTCC power plants in the U.S. obtained from the Energy Information Administration (EIA) data (Figure 3). For each year between 2004 and 2015, the average, maximum and minimum efficiencies (the latter two marked by the error bars) of the top twenty plants are plotted.

Note that cogeneration plants are excluded but some plants are supplementary fired. The efficiency is found by dividing the net generation (in MWh) by the total fuel consumption (in MMBtu, converted to MWh by multiplying with 0.293071) and multiplying by 1.109 (High Heating Value to Lower Heating Value correction).

No correction is made for site ambient and loading conditions. While it is not possible to calculate the capacity factor (CF) of the specific plants in Figure 3, EIA data for all gas-fired combined cycles in the U.S. is plotted in the figure. The latter shows the increasing use of natural gas-fired combined cycle facilities for electric power starting around 2010. Thus, it is not unreasonable to surmise that the 20-unit sample in Figure 3 would also show a similar trend in CF albeit at a much higher value.

ISO base load rating is sticker performance, which is rarely achieved in a real power plant. Site ambient and loading conditions, recoverable and unrecoverable degradation, start-stop cycles, load ramps and planned or unplanned outages all contribute to a performance much below that at new and clean conditions, let alone the rated performance advertised in trade publications.

GE recognized this and came up with a definition to account for those factors known as FlexEfficiency, which is the ratio of profitable annual megawatt-hours to fuel consumption. It is essentially the same as the EIA- 923 efficiencies in Figure 3. Using a predefined operating profile (about 200 starts/year and a mix of base load, part load, and mini- mum turndown hours), GE calculated the typical advanced combined cycle power plant FlexEfficiency as about 56% (or 58.5% for the FlexEfficiency 50 GTCC with a base load efficiency of 61%). Indeed, this number is in line with the upper end of Figure 3.

To estimate the load factor, EIA averages are normalized using the technology curve in Figure 3 (using the formula above) after subtracting one percentage point to account for commercial performance. The result is shown in Figure 4. EIA-923 efficiencies fall within a band of 0.919 to 0.943, which implies a load factor range of 58% to 75%. (Note that the GE FlexEfficieny 50 value is 58.5/61 = 0.959, which corresponds to a load factor of 76% on the curve labeled NEWER.) In fact, in an earlier study, average load factor, when taking into account all elements that impact the performance of a GTCC power plant, has been found to be 74.9% to 77%.

GTCC Carnot factor evolution over the last three decades is shown in Figure 5. With 1,700°C TIT and a Pressure Ratio (PR) of 25, a Carnot factor of 0.85 is required for 65% efficiency (on an ISO base load rating basis —– not consistently and comfortably achievable field performance).

Carnot efficiency, however, cannot be achieved in a realistic heat engine. Available evidence suggests that the limit is around 0.825 —– certainly well below 0.85. There is no reason to expect a technology leap to change this appreciably as long as the industry sticks to the classic Brayton cycle configuration and the brute force approach of ever higher TITs.

Speculation on the future of sticker ratings is not fruitful. One can always come up with pie-in-the-sky developments to enable extreme performances, which are highly likely to remain on paper. Reality in the field demands a metric that accounts for the typical operating regime of a modern GTCC power plant.

Depending on the good fortunes of a given plant site and owner’s economic and financial criteria (e.g., very expensive fuel, access to cheap capital, and so forth), there will always be a bona fide world record holder. However, true state-of-the-art, as demonstrated by the history of coal-fired steam plants in Figure 2, is represented by a more modest benchmark.

Quantitatively and qualitatively, this assertion is borne out by GTCC plant data presented in Figure 3. Even within a sample of 20 plants, the spread in efficiency between the top and bottom plants is about three percentage points.

Furthermore, the average is much closer to the lower end of the band; i.e., only a few plants are truly the best of the best. However one might look at it, the bottom line is unmistakable: Over the last decade, in the largest 60-Hz market in the world, the state-of-the-art in GTCC hovered within 56-57% (LHV) with no clear upward trend (i.e., the upper end of the band for the top 20 units in Figure 3).

On the other hand, such an upward trend is clearly present for the sample mean, which only recently hit 55%. It is not clear whether this is a result of new technology creeping in or higher load factors (or a combination of both).

It thus seems that the industry needs a well-defined mean-effective load factor and corresponding efficiency in addition to the traditional ISO base load rating performance to represent the true capability of GTCC products. This would essentially be a formalized and widely agreed-upon version of GE’s FlexEfficiency. The main advantage would be realism. By corresponding to the situation in the field, it lends itself to more meaningful future projections.

Author

S. C. Gülen is Principal Engineer at Bechtel Infrastructure Power. For more detailed coverage of the concepts discussed herein, see ASME IGTI Turbo Expo 2015 paper GT2015-42077, Étude on Gas Turbine Combined Cycle Power Plant – Next 20 Years, or send an email to scgulen@bechtel.com