«Quadrennial Technology Review 2015 Chapter 6: Innovating Clean Energy Technologies in Advanced Manufacturing Technology Assessments Additive ...»
B: Reciprocating internal combustion engines (RICEs) also present a challenge. Figure 6.D.8 shows the effects of exhaust temperature on overall system thermal efficiencies for a RICE coupled with an RC using various working fluids. The exhaust temperatures vary from hotter (with a 50% brake thermal efficiency baseline engine) to colder (with a 55% brake thermal efficiency stretch engine). Two scenarios are presented: an upper range with higher-efficiency internal RC components and higher exhaust temperatures and a lower range with lower-efficiency RC components and lower exhaust temperatures.
For this combination, efficiency optimization should focus on extracting more piston work, even if doing so reduces the exhaust-gas temperatures and opportunities for waste heat recovery. In some cases, the temperatures may be so low that RCs cannot operate effectively, as seen in Figure 6.D.8 in the steam cycle performance at Texh = 473 K. Cycles not using water might be more advantageous at these scales.
22 Quadrennial Technology Review 2015TA 6.D: Combined Heat and Power
8 Overall system thermal efficiencies (lower heating value basis) for reciprocating internal combustion engine (ICE) with Rankine cycles with various working fluids. Extracting more piston work increases overall efficiency. [Oak Ridge National Laboratory analysis] C: Stirling-cycle engines have been manufactured that operate in the power range (50–60 kW) suitable for waste-heat recovery (WHR). The lower exhaust temperatures from the RICE limit the efficiency of the Stirling engine, and the overall system efficiency improves with more piston work being extracted. This relationship is shown in Figure 6.D.9, where the higher-efficiency RICE leads to lower-temperature exhaust but overall higher combined system efficiency.
23 Quadrennial Technology Review 2015 TA 6.D: Combined Heat and Power Figure 6.D.
9 Thermal efficiency (lower heating value basis) as a function of exhaust temperature of a combined RICE and Stirling-cycle engine system for electrical generation. [Oak Ridge National Laboratory analysis] D: One scheme using a solid oxide fuel cell (SOFC) with a Stirling engine has been proposed52 for a domestic application in the 10 kW range; this system uses a catalytic burner instead of a GT or RICE to oxidize unreacted fuel in the SOFC exit and can operate at atmospheric pressures.
E: The principle behind using an SOFC as a topping cycle is that while it is an efficient electrochemical power generator, its fuel utilization factor can be less than unity, meaning that some fuel (typically, 15%–35%) passes through to the exhaust unless it is recycled. Adding a combustor and work extractor to the exhaust stream uses some of the chemical exergy. Because of their general robustness, GTs typically have been chosen as the bottoming cycle, and one of the limiting factors for GT systems is the turbine inlet temperature (Tinlet).
Typically, in combined-cycle operation, the overall system efficiency increases with Tinlet in a manner shown in Figure 6.D.10. The sensitivity of efficiency with Tinlet (i.e., how much efficiency gain comes with a certain incremental change) is a function of system configuration, such as operating pressures and pressure ratio, use of recuperators or regenerators, and how fuel pressurization and reforming is accounted for in the energy balance.
24 Quadrennial Technology Review 2015 TA 6.D: Combined Heat and Power Figure 6.D.
10 Overall system thermal efficiency (lower heating value basis) for an SOFC-GT hybrid system as a function of turbine inlet temperature. [Oak Ridge National Laboratory analysis] Table 6.D.
7 highlights some SOFC-GT combined-cycle configurations and reported efficiencies; the literature base is much wider. Most of these reported values are numerical model-based and not from experimental studies, so while the range generally can be expected to hold, some unrealistic assumptions or design data should be expected; also, some of these values are reported without plant generation capacity.
7 Efficiencies (lower heating value [LHV] basis) of some SOFC-GT combined cycle approaches to electrical generation.
25 Quadrennial Technology Review 2015 TA 6.D: Combined Heat and Power F: The principal example of an SOFC+RICE system was a planned system by General Electric (GE), using its Jenbacher reciprocating engine. GE has targeted a system efficiency of 65% (lower heating value [LHV]) for cogeneration and 95% for CHP. Generally, RICEs are easier to scale than GT systems because fewer GT systems are currently on the market, but RICEs can be more sensitive to fueling stoichiometries and must be operated with care.
G: One scheme using an SOFC with an RC cycle has been proposed58; this system uses a catalytic burner instead of a GT or RICE to oxidize unreacted fuel in the SOFC exit and can operate at atmospheric pressures.
H: This is the estimation of what is reasonable when the SOFC+RICE system described in (F) is used along with the limits of Rankine-cycle efficiency given expected exhaust properties from the RICE.
I: Triple-cycle systems, typically with SOFC to GT to RC, project efficiencies from 65% up to 78% (LHV) or higher. Most proposed triple-cycle systems include a GT to convert unused fuel from the SOFC, with fuel addition to the GT for proper combustion, and a form of WHR via an RC. The wide range in estimated efficiencies depends on internal thermal optimization of energy flows.
Modeling Methodology Model complexity is typically described in terms of dimensionality, which is a generic description of spatial description and complexity akin to degrees of freedom. For a given flow device such as a turbine or combustor, an imaginary boundary is defined, encompassing the control volume. When all processes within the control volume are lumped and averaged without regard for spatial effects, a zero-dimensional treatment is performed;
in the following discussion, this is referred to as simple modeling. When properties are allowed to vary along a single spatial dimension or zone (for instance, from the inlet to the outlet along the flow path), then a onedimensional treatment is performed. These are examples of low-dimensional modeling. High-dimensional modeling is seen with most computational fluid dynamics simulations, in which a 2-D or 3-D spatial domain is divided into thousands to millions of computational cells and the governing physical modeling equations are solved within each cell.
Generally, the higher the model complexity, the greater the potential for accuracy (with much tuning) but also the higher the cost in modeling effort, sub-model tuning, development time, data validation, and simulation time. With sufficient tuning and validation with carefully crafted experimental data, fairly accurate spatially and temporally resolved predictions of the technology under varying conditions are possible. For scoping analyses such as the present work, low-dimensional treatments are the best means to traverse a range of technologies and configurations. Doing so is a lower-fidelity means than high-dimensional treatments because effects are spatially lumped, time is treated as steady state, and many real processes are not treated in the model. In the present work, simple modeling was used for some systems to gauge a range of performance for given systems to verify that the estimated efficiencies were within the range reported in the literature. The following describes the generic approach employed in the study, except where noted otherwise.
Fluid state properties (e.g., pressure, temperature, enthalpy, entropy, and ratio of specific heats) were obtained by using REFPROP 9.1, a standard software package developed by the National Institute of Standards and Technology.51 REFPROP has interfaces for calling inside of either spreadsheet programs such as Excel or programming environments such as Matlab. Chemistry was simplified as follows: all fuel was assumed to be natural gas, approximated as methane, to compare with standard literature practice and for fuel uniformity.
Combustion was treated as global conversion of fuel and air to carbon dioxide, water vapor with no condensed products, excess oxygen, and nitrogen; because of the state of water vapor, the LHV of the fuel was used for combustion heat (efficiencies were converted to higher heating value [HHV] for reporting as described in the summary). As was typical in the literature, energy required to pressurize the gaseous fuel was neglected (because there are usually different starting pressures and temperatures in practice), and details of any reforming of methane to hydrogen and carbon monoxide for fuel-cell usage were neglected.
26 Quadrennial Technology Review 2015TA 6.D: Combined Heat and Power
Power cycles were constructed by integrating simple models of components at steady state. Where applicable, working-fluid state changes were calculated by using prescribed isentropic efficiencies (for pumps, compressors, and turbines) or effectiveness (for heat exchangers). Flow losses caused by wall friction and geometric effects were neglected, but some flow components had prescribed multiplicative pressure drops defined. Some components were treated as adiabatic (well insulated), with no heat transfer across the boundaries.
The following describes the example formulation of a simple gas-turbine cycle to show the methodology used in GT and RC analysis.
Air starts at the ambient state of pressure P1 and temperature T1. For energy-balance considerations, potentialenergy and other insignificant effects are neglected, and for air flowing through the control volume, its energy flow rate ṁ1, and by conservation of mass, the outlet mass flow rate is ṁ2 = ṁ1. The air enters the compressor, content is described solely by the inlet and outlet specific enthalpy, designated h. The incoming air has a mass which has an isentropic efficiency ηC. By definition, the state change of air from inlet state 1 to outlet state 2 is
defined as follows:
where h1 is the inlet-specific enthalpy, h2 is the outlet-specific enthalpy, and h2s is the outlet-specific enthalpy under an isentropic compression process (the ideality). The compressor component is solved as follows—outlet pressure is defined by a parameter called the pressure ratio rp, which is a key design parameter of the overall
The inlet ratio of specific heats k1 and specific enthalpy h1 for air are obtained using REFPROP. The expected
temperature after compression in an isentropic process is defined as follows:
With P2 and T2s defined and yielding h2s, and using the definition of isentropic efficiency (above), the specific
enthalpy of state 2 is solved as follows:
and with h2 and P2 specified, the temperature T2 is obtained from REFPROP. The required compressor power is
defined as follows:
The fuel stream enters at P2 and T2 and combines with the air at the combustor for a total mass flow rate as follows:
The fuel mass flow rate is a global system parameter that is varied until the overall system electrical output is the target power of 1 MWe. The air flow rate is specified via another control parameter (λ), which is a measure of
excess air and is defined as follows:
27 Quadrennial Technology Review 2015 TA 6.D: Combined Heat and Power where A/F specifies the air-to-fuel ratio. The stoichiometric A/F for methane, on a molar basis, is obtained from
the following global chemical reaction for perfect oxidation of fuel:
combustor exhaust gas temperature entering the turbine, where there is a materials constraint. Knowing ṁF For methane, the mass-based (A/F)stoichiometric is approximately 17.2. GTs run lean, with λ 1, to reduce the and λ, ṁ1 is defined. By conservation of mass on the combustor, the outlet mass flow rate is ṁ4 = ṁ3, and for notational convenience, P3 = P2 and T3 = T2.