Push for Higher Efficiency Turbines Drives Innovation

The National Energy Technology Laboratory (NETL) is developing technology to make coal-burning power plants more efficient and cleaner. Higher efficiencies will mean that less coal is burned, and thus, less greenhouse gases are generated. However, in order to reduce greenhouse gas emissions substantially, the power plants must also be designed to capture, rather than emit, CO₂. To meet these challenges, research is underway on coal gasification, ways to separate the CO₂ from a mixed gas stream (either before or after combustion) so it can be geologically sequestered, oxy-fuel pulverized coal (PC) combustion, and advanced turbines for integrated gasification combined cycle (IGCC) power plants.

Today's typical power plant converts fossil fuel into electricity inefficiently, essentially throwing away over 60 per cent of every unit of fuel we burn. Generally, power plants can be made more efficient by operating at higher temperatures. So, for example, future IGCC power plants will use advanced hydrogen turbines with higher firing-temperature combustors, along with superior materials, better sealing, improved aerodynamics, and advanced mechanical designs. This optimization may include steam cooling of turbine components in conjunction with steam cycle improvements, as well as exhaust gas recirculation (EGR) for NOx control, and pre-combustion capture of CO₂. This will result in an enormous level of thermal and mechanical loading, as turbine inlet temperatures (TITs) will approach 1425-1760 °C (about 2600-3200 °F).

As an alternative to hydrogen-based power generation, DOE is considering the use of oxy-fuel turbines. Inherently designed for carbon capture, the oxy-fuel turbine cycle, fueled with a clean fossil fuel and pure oxygen as the oxidant, offers the potential for similar system efficiencies to the H₂ turbine, with an additional 10 percent CO₂ capture. It also has the benefits of lower NO_x emissions and no CO₂ capture requirement. However, the demand for nearly pure oxygen at a reasonable cost is a major challenge. Furthermore, the turbine for this cycle is far less developed and expected to operate under significantly different working fluid conditions than conventional machines.

Because little is yet understood about either of these two approaches (hydrogen vs. oxy-fuel cycles), it is still unclear as to which will succeed in providing a viable low cost option for meeting U.S. power generation needs. Both will require materials and components that will withstand very high temperatures. To make important contributions that will move development forward, the NETL-ORD Turbine Team is addressing the following key challenges: (1) durability of thermal barrier coatings, (2) accurately determining how well various metal substrates withstand high temperature conditions, and (3) examining the effects of recirculating various levels of exhaust gases on combustion. Each of these research areas will be reviewed in the following sections of this LabNote.

Testing High Temperature Coatings at NETL

The new high efficiency turbines being designed for power plants will require that certain components withstand very high temperatures, from about 1425 to 1760 °C (2600-3200 °F). Currently, there is no commercially available metallic-based material that can maintain a reasonable level of compositional and micro-structural stability for prolonged exposure above 1050 °C (1920 °F). As a result, alloys for high-temperature applications are often protected from surface degra-dation by a thermal barrier coating (TBC) that contains a β-NiAl phase as a primary constituent. These β-containing coatings have a sufficiently high aluminum content to form a continuous scale layer of the slow-growing and highly stable Al₂O₃. Thermally grown oxides can serve as an effective barrier, protecting coated components from the process environment. However, as TBCs are subjected to high temperatures for extended periods of time, rumpling of the bond coat surface can occur, causing induced stress and spallation of the coating, thereby exposing the base metal of the turbine to conditions that will result in quick failure of the turbine.

	Figure 1. The apparatus used to measure pending delamination of the protective layers; evaluations were performed using a spherical tungsten carbide (WC) indenter, a piezoelectric actuator and load cell.

NETL and West Virginia University researchers have developed a micro-indentation technique that can be used to measure the mechanical properties of TBCs applied to alloy coupons at high temperature in high moisture-containing environments to indicate the occurrence of internal debonding and/or spallation. It is based on a load-based multiple loading/partial unloading method that has been successfully used to measure the elastic moduli of various metallic alloys. Yet for application to TBCs, which consists of multiple layers and highly uneven surfaces, micro-indentation evaluations are defined as a surface stiffness response rather than an elastic modulus measurement. These evaluations are performed using a spherical tungsten carbide indenter with a radius of 793.5 μm, combined with a piezoelectric actuator and load cell (Figure 1). Our test results have demonstrated that this micro-indentation technique is non-destructive when small loads (less than 160 N) are applied to the TBC samples (called coupons). [Using scanning electron microscopy, post-mortem cross-sectional micro-structural inspection of areas beneath and around the indented regions revealed no indentation-induced micro-cracking or debonding compared to regions not evaluated with the apparatus. In addition, a comparative thermal cyclic loading of two APS/MCrAlY/RenéN5 TBC coupons was carried out simultaneously, with both the evaluated and non-evaluated sample failing at similar cumulative thermal cycles.]

Figure 2 illustrates how the process works in practice. It shows a coupon exposed to isothermal heating at 1100°C under ambient pressure and composition conditions. The coupon was periodically removed from the furnace (after 85, 185, 300, 400, and 450 cumulative hours of exposure at 1100 °C) for surface stiffness response measurement as well as visual inspection for TBC spallation. After approximately 400 isothermal loading hours, a small visible corner spallation was found between Side 1 and Side 2. Following micro-indentation evaluation, the coupon was subjected to an additional 50 hours of isothermal heating (to a cumulative total of 450 hours), and upon cooling to room temperature, a corner spallation between Side 1 and Side 4 was observed. Again, following micro-indentation evaluation, the coupon was placed back in the furnace for an additional 50 hours of isothermal heating (for a cumulative total of 500 hours). Upon cooling to room temperature, edge spallation along Side 1 was observed. These observations correlated well with sites of measured higher surface stiffness response.

Figure 2. Edge and surface photographs of Coupon B after cumulative 400 thermal heat treatment hours at 1100 deg. C.

Results from the micro-indention measurements are presented to the user as a surface map of the layer stiffness response, see Figure 3. This mapping has been found to correlate very well with subsequent delamination events, as shown in Figure 2.

Figure 3. Surface stiffness mapping The overall conclusion is that this new technique allows evaluation of potential failure of coatings before they result in costly base material failure, and permits early repairs of damaged coating areas.

Turbines Materials

Because coatings alone cannot address the new temperature and environmental extremes being introduced by these advanced cycles, new turbine materials will also be required. The durability performance of existing and next-generation materials is being tested under the unique conditions of oxy-fuel and hydrogen combustion. NETL's Turbine Team, composed of NETL researchers and investigators from NETL's Regional University Alliance, are examining the performance of Ni and Co superalloys as a function of exposure time (1000 to 10,000 hours) to the mixed oxidant environment found in typical oxy-fuel combustion cycles such as the Graz Cycle (approximately 77 percent H2O, 23 percent CO₂, and 0.5 percent O₂). The combustion environment is characterized as a function of O₂ partial pressure (e.g., low, medium, and high, corresponding to 0 percent O₂, 0.2 percent O₂ and 2.0 percent O₂, respectively). Focus at this time is on the intermediate pressure turbine, which has an inlet temperature of approximately 1180 °C. Typical gas turbine superalloys are being tested at conditions that would be expected to occur at various locations within the turbine (approximately 630 to 820°C). This will be done to maximize the practical use of current generation Ni- and Co-base superalloys and extend gas turbine capability to its theoretical limit by optimizing mechanical behavior and environmental resistance to all combustion gases, but especially the high CO₂ content found in typical oxy-fuel combustion cycles (e.g., Graz or CES).

The alloys and coatings of the working components of an existing Siemens SGT-900 gas turbine are shown in Table 1. Five alloys are being evaluated in this test: three nickel base superalloys, and two cobalt base alloys. Internally cooled components are cooled with the working fluid reduced to a lower temperature, but the coated materials are exposed to the actual working fluid (steam + 10 percent CO₂ + 0.2 percent O₂).

The alloys and coatings were exposed at the temperatures in Table 1 for up to 1000 hours. Bare alloy exposures of both Co-base alloys (X-45 and ECY-768) showed tendencies for oxide spallation. The Ni-base alloys (Inconel 738, Inconel 939, and Udimet 520) showed protective parabolic scale growth for all four temperatures, with much more mass gain (i.e., oxidation) at 821 °C than at lower temperatures.

A sample of the results from our testing is shown in Figure 4. Thin chromia scales were found for all the alloys, but with internal oxidation of aluminum for the nickel base alloys. In general, the amount of oxidation on these alloys was relatively low. The nickel base alloys exhibit low corrosion rates, especially at 748 °C and below. The cobalt base alloys are more of a concern, with a tendency towards oxide spallation. The coated specimens showed only a modest amount of growth of the thermally grown oxide and loss of Al from the bond coat. In terms of effective metal loss, internal oxidation is the primary factor of concern for the nickel base alloys.

Table 1. Alloys and approximate metal temperatures for each stage in the SGT-900, as proposed for use as an oxyfuel turbine. Stage Alloy/Coating Approximate Metal Temperature Vane 1 ECY768, Cobalt Base Bond Coat + TBC Internally Cooled 748 °C Blade 1 IN738, Nickel Base Bond Coat + TBC Internally Cooled 821 °C Vane 2 IN939, Nickel Base Bond Coat + TBC Internally Cooled 693 °C Blade 2 IN738, Nickel Base Bond Coat Internally Cooled 748 °C Vane 3 X45, Cobalt Base 748 °C Blade 3 U520, Nickel Base 630 °C

Figure 4. Cross-section of uncoated IN939 after exposure in steam + 10 percent CO2 + 0.2 percent O2 at 821 °C for 1000 hr

Combustion under Variable Exhaust Gas Recirculation Conditions

Exhaust gas recirculation (EGR) of CO₂ will help mitigate the extreme temperatures otherwise present in pure oxygen-fired (oxy-fuel) combustion. NETL research examines how EGR conditions may limit combustion stability (flame stability, combustion dynamics, and emissions) while providing a concentrated stream of CO₂ for post-combustion removal. The NETL Turbine Team employs an integrated approach using lab-scale combustors at NETL, larger-scale atmospheric-pressure combustors at Penn State University, and a high-pressure rig at Virginia Tech.

At NETL, EGR is simulated by diluting a lab-scale burner's fuel-air mixture with varying concentrations of CO₂ and N₂. The dynamic response of the burner is recorded while maintaining a constant exhaust temperature as the level and composition of the diluents is varied. Changes in the inlet composition alter the convective time delay of the dynamic response by varying the flame center of mass. Fuel, air, and diluent are well mixed upstream of the nozzle and injected through a choke-plate to ensure that no gas oscillations in the mixing zone are able to propagate to the flame. The combustion section is a quartz tube (shown in the small figure), 0.56 m in length, which allows self-excited oscillations to occur. A hotwire anemometer placed upstream of the swirler measures both mean and perturbation velocity. A photomultiplier tube is used to image the entire flame heat release rate while an appropriately filtered camera measures the spatially-resolved chemiluminescence. The images are processed to obtain the flame's center of mass.

Recent results from our tests are shown in Figure 5. The plot clearly shows that the flame is elongated (the center of mass moves downstream) by increased levels of diluent. This effect is critical in that it allows for variation of a Strouhal number (defined as the ratio of the acoustic period to the convective time), which is a dimensionless value that is used to analyze oscillating unsteady fluid flow dynamics problems, and allows one to independent control these parameters, and thereby the combustion dynamics.

Finally, Figure 6 shows the amplitude variation with respect to variation in the flame Strouhal number. This figure shows that the amplitude collapses essentially onto a single curve. The significance of this result is that the dynamic behavior of the combustor depends primarily on the convective time delay, and is independent of variations in fuel composition. This suggests that combustor dynamics can be controlled using conventional design practices.