Gas Turbine's Role in Energy Transition

Advanced class heavy-duty industrial gas turbines, in simple or combined cycle, constitute the most efficient thermal power generation technology today (see Tables 1 and 2). Furthermore, utilizing natural gas, even without PCC, they help reduce the carbon footprint of electric power sector (see Table 3). In 2019, per U.S. Energy Information Administration (EIA) statistics, specific CO₂ emissions were 909 lb/MWh, which is 41% of that coal fired generation (2213 lb/MWh). In 2020, according to the U.S. EIA, natural gas accounted for 40% of utility-scale power generation, the largest share among the major sources (cf. 19% by coal). A decade or so ago, the picture was completely reversed, i.e., in favor of coal. The fuel switch from coal to natural gas is the main contributor to achieving 30% reduction CO₂ emissions from the electric power sector in the U.S. in the last decade. To say that a big share of the credit goes to the gas turbine technology is not a stretch at all.

It should be noted that, in Table 3, average efficiency calculation uses 1034 Btu/scf average natural gas HHV (EIA Table A4) and 18,903 Btu/lb (43,968 kJ/kg) average coal HHV (EIA Table A5). Specific CO₂ emission number, calculated using the EIA data, is based on “net electricity generation,” i.e., cogen plants are excluded. For generation and fuel consumption numbers listed in Table 3, theoretical specific CO₂ emission corresponding to 100% CH₄ fuel is 873 lb/MWh, vis-à-vis 2308 lb/MW for coal with the assumption that HHV/LHV ratio is 1.05. (They are listed in parentheses in the table.) In other words, specific CO₂ emissions for natural gas-fired generation are 38% of emissions from coal-fired generation.

The brief survey of state-of-the-art gas turbine landscape in the  Appendix clearly shows the potential of the technology in a sustainable generation portfolio. Before delving into subjects such as hydrogen cofiring and postcombustion carbon capture (from the heat recovery boiler stack gas), let us look at some other possibilities. For example, everything else being the same in Table 3 

• Replacing 10% of coal fired generation with simple cycle advanced class gas turbines (39% HHV) reduces CO₂ emissions by 3.5% (about 57 × 10⁶metric tons or mt).

• Replacing 10% of coal-fired generation with advanced gas turbine combined cycle (GTCC) (55% HHV) reduces CO₂ emissions by 4.3% (about 70 × 10⁶ mt).

• Each 1% (point) increase in average natural gas fleet efficiency reduces CO₂ emissions by 0.8% (about 14 × 10⁶ mt).

• Each 1% (point) increase in average coal fleet efficiency reduces CO₂ emissions by 1.8% (about 30 × 10⁶ mt).

These numbers, without even going into cost and complexity of coal-fired power plants, dramatically illustrate the future direction of thermal power generation. Utilizing existing technology, simply replacing the aging members of the coal fleet with gas turbines, even in simple cycle, makes a significant dent in CO₂ emissions.

There is a large body of literature on carbon capture from flue gases in fossil fuel-fired power plants using chemical solvents. A representative selection is provided for readers interested in the technology and its history [1–7].

Clean stack emissions (from the absorber stack) with different postcombustion capture rates are shown in Table 4. The advantage of GTCC vis-à-vis USPC at the same capture rate is an obvious fact. The interesting fact is that GTCC wins even with a proverbial one hand tied behind its back!

If the GTCC is equipped with a PCC plant with only 85% capture rate, its emissions reduce to 227,650 mt. On the other hand, if the USPC plant is retrofitted with a PCC (same technology) with 95% capture rate, its emissions reduce to 171,100 mt. However, CAPEX and OPEX will approximately double for the USPC plant due to higher CO₂ generation. Considering that the USPC power plant CAPEX is at least five times that of the GTCC to begin with (and higher OPEX), the advantage of about 50,000 metric tons less CO₂ emitted each year is pretty much guaranteed to be not cost-effective.

Another reason for not being too “greedy” about capturing the last gram of CO₂ can be deduced by looking at the big picture. Monthly average CO₂ concentration for 2021 is 419 ppmv (parts per million by volume). Assuming the molecular weight of the atmospheric air as 29 kg/kmol, on a mass basis, CO₂ concentration is 419 × 44/29 = 636 ppmw (parts per million by weight). Taking the total mass of atmospheric air as 5,150,000 Gt, mass of atmospheric of CO₂ is found as (636/10⁶) × (5.15 × 10⁶) = 3274 Gt. (One gigaton (Gt) is equal to one billion metric tons.)

From the U.S. EIA data, using the 2019 total for 2021, probably a conservative assumption, 1.51 Gt total CO₂ is emitted from natural gas and coal-fired electricity generation, which corresponds to 1.51/(5.15 × 10⁶) × 10⁶ = 0.294 ppmw or 0.294 × 29/44 = 0.194 ppmv. The delta calculated above, 50,000 mt or 0.000,05 Gt, is not even a round-off error in the big scheme of things.

The main challenge for postcombustion carbon capture from a GTCC is the low concentration of CO₂ in the flue gas, i.e., about 4%(v). This puts several constraints on the amine absorption process as explained below.

• First, CO₂ partial pressure is low where the flue gas enters the bottom of the absorber and contacts the rich amine—this limits the rate of absorption and rich amine loading practically achievable.

• Second, when considering >90% capture, CO₂ partial pressure in treated gas is very low at the top of the absorber, which limits the rate of absorption and requires an adequately low lean amine loading.

As the energy required to regenerate the amine increases with decreasing amine loading, lower amine loadings increase the reboiler duty requirement for the same quantity of CO₂ captured. Furthermore, there may be practical limits on the level of the capture achievable. Other disadvantages of GTCC flue gas include the high oxygen level, which will accelerate the degradation of the amine.

Advantages of GTCC flue gas, vis-à-vis that from coal-fired plants, are its clean condition (e.g., practically zero sulfur, no mercury, no particulates), low temperature (80–90 °C), and low dew point (around 50 °C or lower). Typical PCC plants require a direct contact cooler (DCC) to accomplish the following tasks:

• Cool the flue gas to a temperature suitable for amine absorption, typically 40 °C

• Collect water condensed from the flue gas

• Remove contaminants, such as SO₂ and particulates from the flue gas

Since the DCC can add as much as 10% to the plant CAPEX, alternatives should be considered when treating GTCC flue gas on the basis that contaminant level is low; it can even be fed to the amine absorber without treatment. Cooling is still required and depending on the sensitivity of the amine to temperature, the following options can be considered:

• Fogging to drop the flue gas temperature to dewpoint (typically 50 °C, which is sufficient for monoethanolamine (MEA) absorption)

• Cooling against water/glycol or air using an exhaust gas heat exchanger. Where cooled below the dew point, the design needs to cater for draining of the condensed water.

Generic solvents can be any commercially available amine but is typically MEA (usually diluted to 30–35%(w)) due to its suitability for this service and a large amount of open-source data from pilot testing is available.

Proprietary solvents consist of one or more amines combined with additives to improve the overall plant performance. Focal points of these improvements are as follows:

1. Reduced energy requirement. The solvent can typically be regenerated with less thermal energy input per ton of CO₂ recovered than a generic amine. This is less important where the fuel cost is low, or the plant operating load is variable.

2. Increased CO₂ loading. The solvent has a higher capacity for CO₂ absorption either due to its chemical structure or by allowing a higher concentration of the amine. This allows a lower solvent circulation rate.

3. Increased stability. Certain proprietary solvents claim to be more resistant to degradation than MEA; if true, this would result in lower amine make up and waste generation. Much longer test durations would be needed to substantiate this difference than have been run with GTCC flue gas to date.

4. Reduced corrosivity. Certain proprietary solvents claim to be less corrosive than MEA; the advantage of this reduced corrosivity may be of limited value since most or all current designs utilize stainless steel for equipment in contact with the solvent.

5. Not surprisingly, as in most cases, benefits are accompanied by potential performance drawbacks, e.g.

6. Slower reaction rate absorbing CO₂, which may require more absorption time and packing volume, lower temperatures requiring more expensive cooling equipment and/or supply cost.

7. More complex chemistry requiring more complex reclaiming, monitoring, and amine quality control.

8. Less desirable emissions, including carcinogenic compounds (e.g., nitrosamines).

9. Several times more expensive than MEA and supply of solvent usually restricted exclusively to the licensor.

Consequently, the decision to use a proprietary solvent needs to be based on a careful evaluation of advantages and disadvantages for the specific project location, constraints, and objectives. As the proprietary solvent will be supplied in conjunction with a licensed process design and license agreement, there will be other design and commercial implications on top of the relative advantages and disadvantages of the solvent itself.

Whether using a generic or proprietary solvent, the required CO₂ capture rate may be dictated by regulatory constraints or open to economic optimization, typically in the range 85–95% capture of CO₂ from the flue gas feed. Main variables to adjust for achieving the required level of CO₂ capture, once the solvent and its concentration are selected, are listed below.

1. Lean amine loading. This is very important for GTCC when targeting CO₂ capture rate > 90% since the CO₂ partial pressure drops to a very low level in treated flue gas and lean amine loading needs to be sufficiently low to achieve this. Lean amine loading is mainly driven by stripper reboiler duty such that higher CO₂ capture requires higher reboiler duty. Also, note that selection of lower stripper operating pressure improves lean amine loading and reduces thermal degradation of solvent in the stripper but increases CO₂ compression horsepower.

2. Structured packing volume. As the absorber diameter is generally set to hydraulic constraints, the packing volume is reflected in the bed height. The bed height needs to be sufficient to achieve the desired rich amine loading. Additional bed height can compensate for a slightly higher lean amine loading and lower stripper reboiler duty. However, further excess bed height in GTCC service is unlikely to be beneficial since CO₂ partial pressure is already low in flue gas feed due to low CO₂ content of GTCC exhaust gas. Thus, rich amine loadings will be limited by the ability of the amine to absorb CO₂. As an example, for MEA this value is found to be around 0.45 mol of CO₂ per mol of MEA.

3. Amine circulation rate. Amine circulation rate is optimized for the CO₂ content in the feed gas. Increasing the amine circulation to reflect increased CO₂ as capture rate increases from 85% to 99% would make sense, but in practice, this requires increasing the reboiler duty. When increasing the CO₂ capture rate, using all the reboiler duty increase to reduce lean amine loading is more effective than to increase amine circulation. Therefore, amine circulation is optimized to feed CO₂ content at say 85% or 90% capture.

Flue gas duct sizing and absorber diameter are fixed by flue gas flow hydraulics. The front end of the carbon capture unit, including the flue gas duct, DCC (if required) and absorber/s, are assumed to be designed for the entire flue gas flow. This part of the design will not change based on the required CO₂ capture rate other than the height of structured packing installed for amine absorption.

The amine circulation rate is fixed based on the flue gas CO₂ content. The back end of the unit, including the lean and rich amine hydraulic and heat transfer systems, the stripper and reclaimer, is designed for similar amine circulation rate. The lean amine loading is achieved by varying the stripper reboiler duty.

The stripper is designed for a lower lean amine loading rate at higher CO₂ capture rates. Typically, as the stripper is much smaller and less costly than the absorber, the packing is given plenty of margin. The only significant change to achieve lower lean amine loading for higher CO₂ capture rates is in stripper reboiler and condenser duties.

Absorber sizing is governed by desired (and practically achievable) rich amine loading. Absorber diameter is selected based on hydraulic considerations. Requisite packing volume is met by selecting the bed height/s, and this will determine the overall absorber height. Typically, the absorber column requires 20 m for bottoms volume, inlet, wash bed, and demisters, and each 10 m packed bed requires further 5 m for supports and distributors. A column with 20 m structured packing is therefore assumed to have a 50 m tan/tan height.

CAPEX increase for higher CO₂ capture rates is driven by the following factors:

• Increased absorber packing volume and column height

• Increased stripper reboiler and condenser size

• Increased CO₂ compressor capacity

OPEX changes are driven by

• Steam consumption—as per increased reboiler duty

• Cooling water consumption—as per increased reboiler duty

• Power consumption—as per increased compressor power for increased CO₂ product plus increased flue gas blower power for added absorber packing bed height/s (higher pressure drop)

• Amine make-up—predominantly driven by unchanged feed and operating conditions so insignificant change, ignored

Operating and maintenance costs are not expected to change significantly. To quantify the impact of capture rate on PCC system design and cost, we have run simulations with Pro-Max 5.0 process simulation software (Bryan Research & Engineering, LLC, Bryan, TX) for a chemical absorption system with 35%(w) MEA solvent. The results are summarized in Table 5.

There are many published studies reporting a broad range of capture cost and performance impact, for coal as well as natural gas fired power plants, based on widely differing assumptions, i.e., capture rate, CO₂ compression pressure, capture technology, bottoming steam cycle design, site ambient conditions, and many more, e.g., see Ref. [3]. Trying to pinpoint a single number, which would be rather misleading anyway, is practically impossible. This contrasts with noncapture GTCC power plant, for which reliable performance (ISO base load rating) and budgetary price information can be found in trade publications, e.g., Gas Turbine World handbooks, published annually. Clearly, at the time of writing, PCC plant design is not at a comparable stage of design, field experience, and construction maturity. Best possible information, on a case-by-case basis, can ideally be obtained in a project-specific FEED (front-end engineering design) study by the EPC (engineering, procurement, and construction) contractor using licensed or generic technology [8,9]. More importantly, piloting of the specific flue gas at a scale that can be used by packing vendors for final selection and sizing of packing is critical. In addition, degradation rates and reclaiming regime for the solvent need to be proven for estimates of solvent usage to be regarded as accurate (for OPEX). This would require at least 1 year of pilot operation.

In order to give a “ballpark” idea, assuming proven chemical absorption technology with a generic solvent, annualized CAPEX of a PCC “block,” including CO₂ compression, for a modern GTCC power plant with a three-pressure, reheat steam bottoming cycle and water-cooled heat sink (including a cooling tower) would be anywhere from $120 to $60 per captured ton of CO₂ (USA average, in 2022 dollars) for power plant capacity factors ranging from about 50% to 90%, i.e., 4000+–7500+ h of annual operation [8]. As far as the performance impact is concerned, similar caveats apply. Especially in retrofit cases, where the PCC block must be designed for integration with an existing power plant with unique site conditions, there is a wide range of net electricity output penalties. Conceptually, for a GTCC designed with an accompanying PCC, one should factor in at least 15% drop in output with respect to noncapture basis. In other words, for a 750 MWe, 62% net LHV (ISO base load) GTCC, including PCC would reduce the net output by 112.5 MWe and the efficiency would drop by 9.3% points.

As discussed above, the innate favorable impact of gas turbines on reducing CO₂ emissions can be further enhanced by retrofitting the GTCC power plants with postcombustion carbon capture. While the authors have refrained from using quantitative CAPEX arguments above (because numbers cited in the literature are unlikely to reflect the complete project cost—to be developed and executed by a competent EPC contractor), there are enough “clues” that GTCC with PCC is highly likely to be the most cost-effective and ready-to-deploy technology, i.e., Technology Readiness Level, TRL, 9, especially with chemical absorption technology using generic solvents) to curb CO₂ emissions. A chart prepared by U.S. DOE's National Energy Technology Laboratory (NETL) provides a relative ranking of different technologies (see Fig. 1).

Hydrogen is widely touted as the holy grail of Decarbonization. The reason for that is simple: its combustion does not generate CO₂. Alas, this is where simplicity pretty much ends. To begin with, H₂ is not an energy resource. It is an energy carrier. Prior to its use as a fuel, it must be produced at the expense of large power consumption, stored, and/or transported. There are significant problems associated with all three phases of the hydrogen value chain. Furthermore, combustion of hydrogen with ambient air generates significant nitrous oxides (NOx), a criteria pollutant. To top it off, burning H₂ in modern Dry-Low NOx (DLN) combustors is extremely difficult due to its unique combustion properties. Until recently, maximum allowable H₂ content in fuel gas was 5%(v). Nowadays, major OEMs offer and guarantee 30%(v) H₂ in their DLN combustors. There are even novel technologies with 60%(v) H₂ capability (claimed). Research and development toward 100%(v) H₂ capability is full speed under way. (In passing, vintage diffusion flame combustors with water/steam injection for NOx control are already 100%(v) H₂ capable but they are completely ignored.)

The “color coding” of hydrogen production technologies is as follows:

• Green Hydrogen is produced by water electrolysis: water is split into hydrogen and oxygen by an electric current and with the help of an electrolyte. If the electricity required for electrolysis comes exclusively from renewable, CO₂-free sources, the entire production process is completely CO₂-free.

• Gray Hydrogen is obtained from fossil fuels. For example, natural gas is converted under heat into hydrogen and CO₂ (steam-methane reforming). Roughly, 9–10 tons of CO₂ are generated to produce one ton of hydrogen from methane.

• Blue Hydrogen is generated CO₂-neutral from fossil fuels. The CO₂ generated during the process is separated, stored, or reused (i.e., Carbon Capture, Utilization, and Storage, CCUS).

There are two commonly used commercially proven technologies to produce H₂: electrolysis and steam-methane reforming (SMR). They will be looked at using numerical examples below. (It should be noted that autothermal reforming (ATR) and gasification are the other two technologies used for H₂ production but at a much smaller scale than SMR. The latter will be covered in Sec. 4.3.) The reader can consult the paper by Dincer and Acar for comparative environmental, technical, financial, and social assessments of 19 possible hydrogen production methods [11].

In electrolysis, electricity is run through water to separate the hydrogen and oxygen atoms. Electricity requirement for H₂ production via electrolysis is very high, around 50–70 kWh/kg. At an average value of 60 kWh/kg, this translates to 216 MWe per kg/s of H₂. Consider a modern gas turbine rated at 300 MWe and 40% of net LHV efficiency. Hydrogen consumption of this gas turbine would amount to about 6 kg/s, which would require 1300 MWe of electricity to produce it from water (consumed at a rate of more than 50,000 gallons per hour)! For a complete picture covering the full range of gas turbine output capacities, refer to the chart in Fig. 2.

Let us assume an advanced class GTCC (60-Hz) in 1 × 1 × 1 configuration rated at 600 MWe with 60% net LHV efficiency. Thus, it consumes 600/0.6 = 1000 MWth of fuel. At 120 MJ/kg LHV for H₂, this comes to 1000/120 × 3600 = 30,000 kg/h H₂ if this gas turbine burns 100% H₂ fuel. Proton exchange membrane technology for green H₂ production requires 55 kWh/kg energy. Ignoring storage, transport, losses, etc., one needs carbon-free (wind, solar, nuke, etc.) surplus power to sustain that single (just one) GTCC. (In other words, the round-trip efficiency, ignoring liquefaction for storage and transportation by any means, e.g., pipeline, ships, etc., and losses, is 0.6/1.65 = 36%. (In order to have a benchmark, as of January 2021, the total installed wind power nameplate generating capacity in the United States is 122.5 GWe.)

Total wind power generation in USA in 2020 was 338,000 GWh –, say, 350,000 GWh. Assuming 5% curtailment (in fact, the average was only 3.4%), 17,500 GWh can be diverted to H₂ production. This can support about two (no typo, just two) 600 MWe GTCC plants for 5000 h/year, i.e., 17,500/(1.65 × 5000) = 2.12. In short, the feasibility of 100% green H₂ for utility scale electric power generation is highly suspect. Let us then look at blending natural gas with hydrogen. In that case, the amount of CO₂ reduction will be a function of the percentage of H₂ in the fuel. The relationship between H₂ content of the fuel gas and CO₂ reduction is as shown in Fig. 3.

Logically, to achieve 50% reduction in CO₂ emissions, based on a heat content basis, requisite fuel gas blend is 50–50 in hydrogen and methane. This is small consolation. A 50–50 H₂–CH₄ blend (in heat content) to reduce CO₂ emissions by 50% would require close to 80%(v) H₂ content in the fuel gas. (At today's state-of-the-art with 30%(v), possible reduction is less than 20%.) In other words, 50% of maximum H₂ flowrate, i.e., 15,000 kg/h for our sample GTCC above, would be requisite. Instead of two 600 MWe GTCC plants for 5000 h/year, four of them could be supported by curtailed wind energy. Without going into lengthy and subject-to-high-uncertainty CAPEX calculations, it should be obvious to an astute reader that, based on the quantitative examples in Sec. 3, retrofitting those four GTCCs with 85% generic solvent PCC would be a more cost-effective solution.

In other words, from one kg of methane and 2.25 kg steam (H₂O), one can make 0.5 kg of H₂ (100% SMR efficiency; real reforming plants make about 0.3 kg H₂ per kg of CH₄, feedstock and fuel). Unfortunately, one also generates 2.75 kg of CO₂. This is the so-called gray H₂. Even more unfortunately, if one accounts for the fuel burn in the reformer, CO₂ production can reach (round number) 5 kg (per kg of CH₄ feedstock). Let us do a tradeoff analysis with CH₄ LHV = 50 MJ/kg and H₂ LHV = 120 MJ/kg. Assuming 60% net LHV GTCC efficiency and ignoring all other stuff (H₂ storage, transport, losses, etc.),

• CH₄ electricity generation: 1 × 50 × 0.6 = 30 MJe/kg

• H₂ electricity generation: 0.5 × 120 × 0.6 = 36 MJe/kg

• CH₄ CO₂ generation (GTCC stack): 2.75 kg/s or 329.9 kg/MWh

• H₂ CO₂ generation (SMR stacks): 5.0 kg/s or 5 × 3600/36 = 500 kg/MWh

It is eminently clear that gray H₂ is a futile proposition. Now, however, if one adds 90% carbon capture (CC) to the SMR, the picture changes, i.e., CO₂ generation is reduced to only 50 kg/MWh. This is the premise behind the blue H₂. However, the thing is that one can match that with only 85% PCC added to the GTCC, i.e., 0.15 × 329.9 = 49.5 kg/MWh. Let us now do some thinking. Which one makes more sense:

• GTCC + PCC (85%) for < 30 MJe/kg electricity and 49.5 kg/MWh CO₂ emission, or

• SMR + CC (90%) (+ H₂ storage + H₂ compression/piping) + GTCC for < 36 MJe/kg electricity and > 50 kg/MWh CO₂ emission?

(Note that, everything else being the same, 85% capture has less parasitic power consumption vis-à-vis 90% capture. On the other hand, at least some of the SMR CO₂ streams are more amenable to capture, e.g., higher concentration and higher pressure) It is left to the reader to draw his or her own conclusions.

A third possibility is gasification of coal, refinery residue, or other problematic hydrocarbons. Unlike the two prior examples, in this case, the feedstock, in contrast to surplus electric power plus water or natural gas (mostly methane, CH₄), is typically not a good candidate for “clean” electric power generation. If the CAPEX and OPEX numbers can be made to work out, gasification of a “not green” feedstock such as coal, petcoke, and refinery residues (or biomass for that matter) with carbon capture and storage to produce blue H₂ can be a viable way to clean, sustainable electric power generation.

As an example, let us look at gray or blue H₂ from gasification of a “nasty” hydrocarbon feedstock. The example assumes a hydrogen plant defined as follows [12]:

• Feedstock is visbreaker residue (20,420 kg/h),¹ whose composition is about 85%(w) carbon (C), 10%(w) hydrogen (H), and 4% sulfur (S) with vanadium, nickel, iron, sodium, and calcium in the mix

• The feedstock and oxygen are sent to a partial oxidation reactor (flow ratio is roughly 1:1), to which O₂ is supplied from the cryogenic air separating unit (ASU)

• After raw gas shift, a Selexol™ process is used to separate H₂S and CO₂ (42,988 kg/h) from the syngas (H₂S is converted to sulfur product in a Claus unit)

• Clean gas is sent to a pressure swing adsorption block to generate hydrogen (4464 kg/h)

• Hydrogen produced by this process can be used in a GTCC to generate electric power. Assuming net 60% LHV efficiency, specific output is calculated as 72 MJe/kg. For a fair power accounting, one should subtract from this number ASU, Claus unit, and gasifier power consumption. Let us assume that the net specific output thus drops to 60 MJe/kg.

Under normal circumstances, visbreaker residue and similar refinery byproducts are not suitable as gas turbine fuels. In the old days, such ash-bearing fuels were utilized for marine, refinery, and oil-field applications. Significant pretreatment (e.g., washing) and use of additives (inhibitors) were requisite to prevent turbine hot gas path deterioration. The reader is referred to the GE white paper by Kaufman for details [13]. A further difficulty presents itself in the heat recovery boiler (HRB) with exhaust gas containing high levels of sulfur. There is only one GTCC power plant in the world designed to burn refinery bottoms for power and steam cogeneration: Kalaeloa Cogeneration Plant in Hawaii [14]. Expectedly, its performance was not stellar, and the operators had a tough time with extreme fouling of HRB tube banks due to sulfuric acid formation.

Net GTCC output with 100% H₂, using the gray H₂ generated by gasification, is calculated as (4464/3600) × 60 = 74.4 MWe with 578 kg/MWh CO₂ but pretty much no SOx. With the addition of CC to reduce CO₂ emissions to, say, 57.8 kg/MWh (90% capture), the feasibility of this blue H₂ option depends on the $/MW CAPEX (when everything is thrown in, i.e., from the raw feedstock to the final megawatt-hours of electricity delivered to the grid), OPEX, and $/MWh LCOE (levelized cost of electricity).

One should also mention gasification of biomass or coal plus biomass with carbon capture for net carbon-negative power and/or hydrogen production. For a comprehensive study of this technology, the reader is referred to the NETL report from 2012 [15]. The study covered 47 cases with varying emission targets, i.e., 0, 350, 800, and 1100 lb CO₂/net-MWh. The report concluded that it was theoretically possible to achieve net zero life cycle emissions at demonstrated levels of biomass (switchgrass) cofiring (30% by weight) and less than 90% CO₂ capture. It was, however, not possible to achieve net zero life cycle emissions in a 100% switchgrass-fed plant without CO₂ capture. The greenhouse gas (GHG) emissions associated with growing, harvesting, and transporting the switchgrass required about 14% of the power plant carbon emissions to be captured.

Compressed Air Energy Storage (CAES), with the exception Pumped Hydro-Storage (PHS), is the only proven (i.e., TRL 9) large-scale, long-duration energy storage technology utilizing turbomachinery, specifically, reheat gas turbines [16]. At the time of writing, there are only two CAES plants in the world: the 290 MW plant in Huntorf, Germany (commissioned in 1978) and the 110 MW plant in McIntosh, Alabama, which was commissioned in 1991. The underground storage in either plant is a solution-mined salt cavern (the Huntorf plant has two caverns).² Alternatives have been considered for many proposed CAES projects in the four-decade period since the commissioning of the Huntorf plant. They include limestone mines, depleted natural gas reservoirs, aquifers (water-filled underground reservoirs), and porous rock formations. Each have their advantages and disadvantages as well as CAPEX implications. High CAPEX and inability of developers to secure funds have prevented the technology from reaching a mature commercial readiness level. The primary deficiency—as perceived by the lenders and insurers—was the lack of firmness in projected revenue stream. This might change with the right regulatory and/or market incentives.

In addition to the conventional CAES technology exemplified by the two plants cited above, several notable variations proposed have been proposed recently. One of them is the adiabatic CAES with storage of heat rejected by compressor inter- and aftercoolers during charging for use during discharging and power generation [17]. Another improvement includes hydrostatically compensated air storage so that the reservoir pressure stays constant during discharge operation [18]. Probably, the biggest technology hurdle in CAES is either finding a suitable storage location (e.g., a depleted natural gas reservoir) or making one (e.g., solution mining a salt cavern). An alternative solution is provided by cryogenic or liquid air energy storage (CES or LAES), where air is cryogenically liquefied for storage in tanks [19]. Typically, this technology includes thermal energy storage and avoids combustion during the discharge phase (power generation).

Gas turbines can also be used for utilizing solid or liquid fuels in lieu of conventional thermal power plants with boiler and steam turbine. (Strictly speaking, steam is a dense gas, and the steam turbine is thus a particular gas turbine variant. In other words, a conventional thermal power plant can be characterized as a huge closed-cycle gas turbine.) Around the same time as that of commissioning of the BBC Neuchâtel gas turbine, in 1939, another Swiss company, Escher-Wyss, built a 2 MWe test installation in their factory in Zurich. The power cycle, with air as the working fluid in a closed-loop, was named after its inventors, J. Ackeret and C. Keller, as the AK cycle. Eventually, between 1950 and 1981, several experimental and commercial closed-cycle gas turbine power plants utilizing coal and steel mill off-gases as fuels were built in Europe, Japan, and the USA [20].

Nevertheless, the inability of closed-cycle machines to match ever-increasing turbine inlet temperatures (TIT), simplicity, and agility of industrial gas turbines, which are similar to jet engines in construction and thermal cycle, ended their run by early seventies. In a nutshell, a cost-effective closed-cycle gas turbine design is out of question at high temperatures requisite for efficiencies in the same league as a modern GTCC. However, they can utilize solid fuels such as coal and biomass with efficiencies comparable to conventional thermal power plants, especially with supercritical CO₂ as the working fluid.

There are many potential working fluids for closed-cycle power plants [21]. The last decade has seen a significant R&D activity in power cycles using supercritical CO₂ [22–29]. In principle, a closed-cycle gas turbine with supercritical CO₂ as its working fluid (SCO₂) is an excellent bottoming cycle to complement a topping cycle with low-grade exhaust energy (e.g., an aeroderivative gas turbine or reciprocating gas-fired engine). It is also a good fit to advanced nuclear reactors, e.g., Generation IV reactors such as LMR (Liquid Metal Reactor). Particularly, a sodium-cooled fast-breeder LMR is ideally suitable to SCO₂ Brayton cycle with ∼500 °C at the exit of the intermediate heat exchanger [26,27]. Similarly, an SCO₂ turbine, small enough to be put at the top of a solar tower along with a solar receiver, is an intriguing option for concentrated solar power [28].

Net output and efficiency of an SCO₂ power plant as a function of TIT are plotted in Fig. 4 (same cycle pressure ratio, i.e., 3.33:1 at the turbine). At TIT of 850 °C (1562 °F), net efficiency is 54%. Note that the cycle model used in this sample calculation does not account for turbine cooling. Depending on the materials used in turbine casing and flow path (including the turbine rotor), some cooling is likely to be required with detrimental impact on cycle efficiency. The other design challenge is associated with material selection and sealing for heat exchanger, piping, and the inlet (throttle) valve (control and stop functions). In essence, temperatures beyond 600 °C—especially at pressures pushing 300 bar—are challenging from hardware design, development, and demonstration perspectives. Difficulties encountered in development of piping, valves, and turbine materials for ultrasupercritical fossil fuel-fired power plant technology provide ample cautionary tales in that regard.

The Allam Cycle is essentially a semiclosed cycle gas turbine with oxy-combustion wherein CO₂ constitutes (nominally) 95% of the fluid flow in the combustor (by mass) with the rest, 5%, made up by oxygen and fuel. Oxygen for combustion is generated by a cryogenic ASU. Carbon dioxide generated by the combustion is taken away from the cycle at the CO₂ pump discharge to maintain the cycle mass balance (hence semiclosed). The resulting combustion product is roughly 90% CO₂ and the ASU parasitic power consumption is minimized by the lower O₂ requirement. The claimed net LHV efficiency of the Allam cycle is nearly 59%. A detailed plant performance analysis based on the Allam cycle is provided by Sifat and Haseli in a recent paper [30]. The authors in the cited paper compare the claimed performance from the 2016 paper by Allam et al. [31] with their own model predictions. No information is available on the details of the cycle turbine (provided by Toshiba), e.g., number of stages, stage efficiencies, and cooling flows. Sifat and Haseli's simple model assumed an uncooled expander with 90% efficiency. Compressor and pump efficiencies were taken as 85% and 75%, respectively. Even with such simple yet optimistic assumptions, the authors could only predict 55% efficiency with the same cycle parameters reported by the original developers. Using the assumptions from the MS thesis of Manso [32], the authors could come up with only 51.8% (lower turbine efficiency). A detailed model by the first author, using the same design assumptions, could only reproduce 52.1% net thermal efficiency (open loop, water-cooled heat sink, accounting for O₂ and fuel gas booster compressors). Oxygen, however, must be manufactured in a very capital and electric power-intensive plant, namely, the cryogenic ASU. On top of that, in many places in the world, returning circulating cooling water carrying a significant amount of heat at 30 °C cannot be dumped into a lake, river, or ocean; not anymore. This will require an electric fan driven heat rejection system such as a dry cooling tower or bank of fin-fan coolers. Finally, one must account for miscellaneous power users in the plant (0.1% of gross output) and generator step-up transformers in the switchyard (0.35% of gross output). With all these additions, the realistic power plant efficiency becomes 43.5%—competitive with the best coal fired technology without PCC. (Modeling and calculation details will be available in the upcoming book (in 2022) by the first author [33].) Even ignoring all these parasitic losses, at 52% net, this technology, barely at a TRL of 6 or 7, can only match an advanced class GTCC with PCC (see the end of Sec. 3), the most mature and cost-effective technology available today. (In that regard, also see the U.S. DOE technology chart in Fig. 1 above.)

Gas turbine technology is an ideal complement to “green” electricity initiatives in the era of Energy Transition with the prime focus on Decarbonization. Gas turbines in simple or combined cycle can form the “bridge” in transition from fossil fuels to renewable, zero-carbon resources by replacing coal fired generation and help firm the generation portfolio with higher solar and wind content. For the sake of full disclosure, however, we should also say a few words on the drawbacks.

While superior to coal and oil in terms of emissions, CO₂ as well as criteria pollutants, natural gas is a hydrocarbon fuel comprising mainly methane (CH₄), which is a GHG more potent than CO₂ in terms of its “warming potential” (i.e., trapping radiation). It constitutes about 10% of anthropogenic GHG emissions in the USA (in 2019 per EPA, which, per some researchers, understates the magnitude of emissions from the oil and gas industry, i.e., 8 × 10⁶ mt instead of 13 × 10⁶). (Luckily, methane has a shorter lifetime in the atmosphere.) While not considered in GHG emissions calculations herein, one cannot ignore the fact that methane is emitted during production, processing, storage, transport, and distribution of natural gas (as well as coal and oil for that matter). (In all fairness, methane emissions also result from livestock and other agricultural practices, land use and by the decay of organic waste in municipal solid waste landfills.) There are claims (not definitive proof) that methane emissions pretty much canceled out the gains from the CO₂ emissions reduction by coal-to-natural gas switch mentioned earlier. Indeed, per U.S. Environmental Protection Agency (EPA), in 2019, methane emissions from the oil and gas sector amounted to 0.197 Gt of CO₂ equivalent.³ Adding this to 0.56 Gt of CO₂ emitted by natural gas fired electricity generation in 2019, the total comes closer to coal fired number, i.e., 0.76 Gt vis-à-vis 0.95. It is thus imperative to introduce strict measures for methane leak prevention from the activities in the natural gas value chain. Unfortunately, this is easier said than done due to the difficulty associated with precise identification of sources and quantification (measurement) of associated leakage.

One concern that is mentioned in the literature is the direct and indirect CO₂ emissions during production of MEA, which is produced via an exothermic reaction (i.e., no catalyst is required) of aqueous ammonia and ethylene oxide (EO)—see the paper by Luis [34]. This concern is unwarranted. CO₂ from ammonia production is highly concentrated and therefore readily and inexpensively compressed and sequestered. Furthermore, most of the CO₂ produced is combined with a side stream of ammonia to produce urea, a salable product. In any event, MEA captures many times its weight in CO₂ over its approximately 1-year life in the system. As an example, if net capture per cycle is 0.25 kg/kg and cycle time is 1 h, then in 28 h roughly seven kg of CO₂ is offset. That leaves more than 8700 h for positive capture balance.

In this paper, we investigated the various ways of using gas turbines in playing a key role in Energy Transition or, more specifically, in Decarbonization. Simply due to favorable characteristics of natural gas, relative to other hydrocarbon fuels, and the technology itself, i.e., high thermal efficiency, gas turbines, especially in combined cycle configuration, already made a significant dent in CO₂ emissions from the electricity sector in the U.S. by replacing coal-fired generation. This trend is certainly going to continue—as it should.

As far as hydrogen is concerned, our calculations do not support the hype created around this “magic” fuel. The biggest problem is its availability. Hydrogen is not a “resource” that can be mined or extracted like other hydrocarbon fuels. It must be manufactured. Green H₂ production by electrolysis consumes so much power that its overall efficiency is barely 25% if one counts everything between the end points of the total value chain. In other words, 1 MWe of, say, renewable or other primary resource spent in manufacturing H₂ translates into 250 kWe when it is burned to generate electric power in a GTCC power plant. Exorbitantly large resources would have to be dedicated to green H₂ manufacturing to sustain a reasonable number of GTCC power plants (if they are to burn 100% H₂—not a possibility yet with DLN combustors). Co-firing, say, 30%(v) H₂ with natural gas (NG) is a proverbial drop in the bucket. It is no more than a “cosmetic” attempt to look “green” without making a meaningful dent in the GHG emissions problem and help reach the Paris Agreement goals.

Similar arguments can be made for H₂ produced via SMR. Without capturing the CO₂ generated during the SMR process, gray H₂ is an exercise in futility. It is more efficient and cost-effective to utilize the original feedstock, i.e., natural gas, directly in a GTCC. Deploying the CO₂ capture technology in the SMR plant, i.e., manufacturing the so-called blue H₂, is not a meaningful option, either. If the idea is to build new SMR plants dedicated to gray or blue H₂ production, it is highly unlikely to be more cost-effective than simply burning natural gas in a GTCC with or without PCC. One does not even have to do a full-blown CAPEX estimate to be able to see that. Utilizing the H₂ produced in existing SMR plants can be an option but the same argument made above for H₂–NG cofiring is equally valid. It is just a cosmetic exercise.

On the other hand, utilizing gasification with otherwise unsuitable-for-GTCC feedstocks such as coal, petcoke, refinery residues, and all other types of “nasty” hydrocarbons or even biomass for H₂ production might be a feasible option. A diligent CAPEX/OPEX analysis with a proverbial “sharp pencil,” ideally in a full-blown FEED study by an experienced EPC contractor in collaboration with OEMs, is requisite to prove the said feasibility. If this route to H₂ production turns out to be viable indeed, it would constitute one more route to “clean” electric power generation by making use of ample hydrocarbon resources, which otherwise do not belong to the era of Energy Transition.

The current technology status of gas turbine technology did not happen overnight. It started with Hans von Ohain's hydrogen-fired radial turbojet HeS-1 in his laboratory (1936–1937) and BBC's 4-MW gas turbine in Neuchâtel, Switzerland, for electric power generation (1939). The first gas turbine as we know it today, i.e., the core of Jumo-004 turbojet engine, went into serial production and active war service (in Messerschmitt Me 262 interceptor) in 1944, near the end of WWII. In the ensuing eight decades, thermal efficiency of gas turbines progressed from barely 15% (net LHV) to well above 40% as shown in Fig. 5. While the basic engine configuration stayed unchanged, the technology development took place primarily in the engine thermodynamic (Brayton) cycle, i.e., cycle pressure ratio (PR) and TIT. This is illustrated by the “class hierarchy” of heavy-duty industrial gas turbines (commonly known as “frame machines” to distinguish them from the aeroderivatives), i.e.,

• E Class – 1300 °C TIT (PR = 12:1)

• F Class – 1400 °C TIT (PR = 15:1)

• G Class – 1500 °C TIT (PR = 20:1)

• H System (discontinued) – 1500 °C TIT (PR = 23:1)

• H Class – 1500 °C TIT (PR = 21:1)

• J/JAC Class – 1600 °C TIT (PR = 23:1)

• HA Class – 1600 °C TIT (PR = 23:1)

• HL Class – 1600 °C TIT (PR = 24:1)

Note that G, J, and H-System have steam-cooled hot gas path (HGP, i.e., the turbine) parts. Furthermore, listed TIT values are introductory; HA, J/JAC, and HL class TITs are now most likely 1650+° C (PR of 25:1 for J/JAC). For the direct relationship between cycle PR and TIT and the underlying thermodynamics, the reader is referred to the monograph by Gülen [36]. Early developers of gas turbines and jet engines already knew that higher cycle PR and TIT were requisite for higher thermal efficiency. The missing ingredient was HGP component (blades, vanes, disks) materials that could withstand such high temperatures without melting away in seconds. Eventually, development of nickel-based superalloys, thermal barrier coatings (TBC), directionally solidified (DS) and single crystal (SX) casting technologies, and increasingly intricate film cooling techniques brought the technology to 1700 °C TIT and 25:1 pressure ratio [36].

Not surprisingly, the advance in cycle thermodynamics came with its own bag and baggage, i.e., excessive CO and NO_x (a criteria pollutant) production in the combustor, secondary (cooling) and combustion air flow management problems due to elevated compressed air temperatures (result of higher cycle pressure ratio), increasing HGP cooling requirements, and larger compressors for high airflows requisite for gas turbine outputs pushing beyond 500 MW (50-Hz units) [36]. Solving all these problems required significant advances in the following areas, in addition to those mentioned above, especially in the last three decades (starting with U.S. Department of Energy's (DOE) Advanced Turbine Systems program in the 1990s): premix or Dry-Low NOx (DLN) combustion, three-dimensional (3D) aerodynamics, advanced seals, tight clearances, and model-based adaptive control. Recently, two emerging areas of technology can be added to the mix as well, i.e., additive manufacturing (colloquially known as 3D printing) and digital twins.

In specific hardware terms, modern gas turbines share the following characteristics:

• Single shaft, two-bearing design

• Can-annular DLN combustor

• Four-stage turbine (hot gas path)

• 12–14 stage compressor with 3D airfoils and variable guide vanes (VGVs)

In addition to these common features, OEMs have their own proprietary technologies such as hydraulic (active) clearance control, single tie-bolt rotor design (same as in 1940s vintage Jumo-004, combining individual disks with Hirth serrations), axial fuel staging, rotor cooling air cooling, enhanced air cooling, and steam cooling, to name a few.

Apart from the conventional gas turbine architecture described above, one should also mention two noteworthy variants: reheat (sequential combustion) gas turbine and steam-cooled H-System (now defunct). The former is the practical implementation of a fundamental Brayton cycle enhancement, i.e., reheat, whereas the latter comprises an “externally cooled” hot gas path (mostly). Reheat gas turbines, GT24 (60-Hz) and GT26 (50-Hz, former ABB/Alstom, now owned by GE (the 60-Hz version) and Ansaldo (the 50-Hz version) are still around, whereas the H-System was eventually dropped by GE due to its complexity and cost even though it operated quite successfully in the field.

Frame machines are rarely deployed in simple cycle configuration. First introduced in early 70 s, combined cycle power plants with a steam “bottoming” Rankine cycle, comprising a HRB and steam turbine generator (STG), came into their own in 1990s and presently form the backbone of utility-scale electric power generation. For an in-depth coverage of the technology, the reader is referred to Ref. [36]. The development of the technology in past four decades closely follows that of the simple cycle technology in Fig. 5 and is summarized in Fig. 6.

Field performance of combined cycle plants can be examined in two categories: (i) field test performance with diligently specified site and ambient boundary conditions and “new and clean” equipment and (ii) annual (average) performance during normal commercial operation. The former is typically hard to come by. Nevertheless, lately, OEMs have been keen to advertise their products' “world record” combined cycle performances, as measured in the field, so that we are able to get a feel for the “actual” technology performance as opposed to the “rating” numbers from the trade publications. Notable examples are:

• Irsching (Siemens, 50-Hz, H Class, 2011) 60.75% net LHV (0.6 in. Hg backpressure, cooling water from the Danube)

• Bouchain (GE, 50-Hz, HA Class, 2016) 62.22% net LHV (old coal plant's natural-draft cooling tower, auxiliary power is excluded)

• Nishi Nagoya (GE, 60-Hz, HA Class, 2018) 63.08% gross LHV (once-through, open loop seawater cooling)

• Lausward (Siemens, 50-Hz, H Class, 2015) 61.5% net LHV (cogeneration plant)

The second type of information is readily available online, i.e., EIA-923 Monthly Generation and Fuel Consumption Time Series data. A selection of GTCC power plant data from 2020 Form EIA-923 (final) is summarized in Table 6. Individual plants are identified by their ID numbers. (The EIA form does not include gas/steam turbine OEMs and models. That information is researched and added by the authors.) The list covers products from the major OEMs (identified as A, B, and C). Instead of the exact gas turbine model, only the machine class is identified. The original form itself can be accessed by the reader from the EIA website.⁴ Unfortunately, the data do not include plant capacity factors. For some plants, there is scant publicly available information to enable an approximate calculation. For instance, GTCC power plant #60345 in Table 6 is a 3 × 3 × 1 H class GTCC with three 337 MW gas turbines and a 550 MW steam turbine. With 2% auxiliary load assumption, net plant output is estimated as 1530 MW. Thus, without including degradation, loading, and ambient effects (unknown), plant capacity factor is found as CF = 9,463,975/(8760 × 1530) = 0.706 or 71%. Since this calculation is subject to high uncertainty due to missing/scant information on plant condition and operation, it is left out from the table. However, a reader with more concrete information can easily do this calculation more accurately using EIA-923 data.

As the data in Table 6 show, the majority of the GTCC power plants in the USA have supplementary firing for power augmentation, especially on hot summer days when power demand and prices peak. While the data lists fuel consumption in the HRB duct burners separately, there is no other information to derive a meaningful unfired efficiency from the numbers. For three plants in the list, #55451, #59812, and #60122, duct burner fuel consumption is 38.5%, 34.6%, and 36.4%, respectively, of total plant fuel consumption, i.e., they are heavily duct-fired.

Data presented above provide a synopsis of GTCC technology state-of-the-art at the time of writing, i.e., at the beginning of the third decade of the twenty-first century:

• In terms of ISO baseload rating (analogous to the sticker performance of automobiles), most advanced GTCC power plants push 900 MWe output with close to the 65% net LHV efficiency goal post (Table 2).

• In the field, the same technology, depending on site ambient and loading (dispatch) conditions, wear, and tear (degradation), existence of HRB duct firing for power augmentation, and other factors, registers in mid-50 s (% LHV) on an annual average basis (Table 6).

• U.S. natural gas fired fleet's 2019 performance (including all plant types and vintages) is about 50% LHV (from Table 3, converting 45.1% HHV to LHV). A typical selection (in Table 5) of later vintage GTCCs (fired and unfired) has an average of 49.3% HHV and 54.7% LHV.

• It is interesting to note that the latest gas turbine technology at 40+% net LHV (ISO base load rating, see Table 1) is decidedly superior to U.S. coal-fired fleet's 2019 performance (32.4% HHV in Table 3, probably around 35% in LHV).