Article pubs.acs.org/EF
Investigation about Temperature Effects on Soot Mechanisms Using a Phenomenological Soot Model of Real Biodiesel Xiaojie Bi,*,† Xinqi Qiao,† and Chia-fon F. Lee‡,§ †
Key Laboratory for Power Machinery and Engineering of Ministry of Education, Shanghai Jiao Tong University, Shanghai 200240, People’s Republic of China ‡ Department of Mechanical Science and Engineering, University of Illinois at Urbana−Champaign, Urbana, Illinois 61801, United States § Center for Combustion Energy and State Key Laboratory of Automotive Safety and Energy, Tsinghua University, Beijing 100084, People’s Republic of China ABSTRACT: A phenomenological soot model of real biodiesel was proposed to investigate the effects of initial ambient temperatures on combustion and soot emission characteristics of soybean biodiesel. Validation experiments were conducted in an optically accessible constant volume chamber under four difference initial ambient temperatures: 1000, 900, 800, and 700 K. Good agreement was observed in the comparison of time-related soot measurement and prediction. Results indicated that ignition delay prolonged with the decrease of the initial ambient temperature. The heat release rate demonstrated the transition from mixing controlled combustion at a high ambient temperature to premixed dominate combustion mode at a low ambient temperature. Although the soot formation and oxidation mechanisms were both suppressed, biodiesel showed less soot tendency at a lower ambient temperature. Temporal and spatial distribution pictures indicated that the drop in ambient temperature did not cool the combustion temperature. The reduction of the soot mass concentration with the decrease of the initial temperature was caused by the shrinked total area of a local high equivalence ratio, in which soot usually generated fast. At 700 K initial ambient temperature, soot emissions were almost negligible; therefore, clean combustion might be achieved at super low initial temperature operation conditions.
1. INTRODUCTION Under the pressure of an aggravating global energy crisis and environmental problems, biodiesel, as a kind of renewable alternative fuel, has attracted more and more attention. Commonly, biodiesel refers to fatty acid methyl or ethyl esters that derived from a variety of renewable vegetable oils, animal fats,1 and more recently, algae.2 Similar to fossil diesel, biodiesel is compatible with current fuel injection infrastructure and can be applied to current diesel engines without considerable modifications.3−5 On the basis of the comparative experiments conducted in compressed ignition engines, diesel and biodiesel fuels indicated similar combustion characteristics and almost identical power output. Life-cycle analysis of biodiesel also pointed out that biodiesel use can mitigate CO2 emissions because CO2 is consumed in the growth of requisite feedstock for biodiesel manufacture.6 Soot, a major component of the particulate matter (PM), is one of the key air pollutants emitted by diesel engines.7−9 Previous research proved that those soot nuclei with a radius between 0.1 and 0.5 μm are capable of directly depositing in the lungs, resulting in serious health issues.10,11 In addition to the emission challenge, soot formation in diesel engines can also influence engine performance and have feedback effects on incylinder combustion and emission characteristics. Soot, once presenting within a flame, plays an important role in radiative heat transfer:12,13 soot produces broadband incandescent radiation, which typically dominates over the narrow-band radiation from intermolecular processes. Therefore, soot appearance within the flame would enhance the flame © 2013 American Chemical Society
emissivity and, hence, increase the radiative heat loss. The combustion efficiency then turns to decrease. As a kind of oxygenated fuel, biodiesel, whether neat or in blends, showed a natural PM benefit in diesel engine tests. Previous investigations proved that soot emissions declined steadily as the biodiesel concentration increased in blend fuel.14−17 The amount of soot reduction had a direct link with the total amount of biodiesel added to the diesel fuel, and each kind of biodiesel had approximately the same effectiveness in reducing soot emissions. Therefore, it is reasonable to believe that soot or PM emissions could be reduced by modifying the diesel fuel composition. Besides fuel composition, there is other parameters, such as the temperature, pressure, equivalence ratio, engine design, and operation parameters, involved in soot evolution, among which the temperature plays the most important role.7 Soot and temperature have an inherent coupled dependence: temperature depends upon the soot concentration because of heat transfer through radiation, and soot depends upon the temperature because of the chemical and physical processes controlled by the temperature. Therefore, a detailed understanding of soot must depend upon the in-depth analysis of the temperature. Multi-dimensional simulation is developed and applied to obtain an in-depth understanding about combustion and soot Received: June 27, 2013 Revised: August 18, 2013 Published: August 21, 2013 5320
dx.doi.org/10.1021/ef401208b | Energy Fuels 2013, 27, 5320−5331
Energy & Fuels
Article
reaction mechanisms. Over the decades, great efforts have been taken to develop biodiesel combustion and emission models. Fisher et al.18 created a chemical kinetic model for methyl butanoate and normal alkanes up to n-butane, which consists of 279 species and 1259 reactions. On the basis of the experimental measurement in a jet-stirred reactor, Dayma et al.19 created a kinetic mechanism for methyl hexanoate (MHEX). Through the validation against their measurements, the model by Dayma et al. successfully predicted the reactivity in the cool flame, negative temperature coefficient, and hightemperature oxidation regions. On the other hand, Herbinet et al.20 developed a chemical kinetic model for methyl decanoate, which assembled mechanisms of n-heptane,21 iso-octane,22 and methyl butanoate with both low- and high-temperature combustion chemistry. Biet et al.23 developed a chemical kinetics model for C9−C17 methyl esters and, subsequently, validated it with experimental data for methyl palmitate conversion in a jet-stirred reactor. The agreement in the comparison between computational results and measurements indicated that the model was able to predict the effect of the temperature on the conversion percentage. However, these kinetic models of simpler surrogate molecules do not necessarily represent the complete features of biodiesel combustion. Also, because these detailed kinetics models usually included hundreds of chemical reactions, they were super time-consuming in diesel engine simulations because of the limitation of computer capacity and the complex nature of turbulence, thermochemistry, and multiphase systems. In history, an empirical model, such as Hiroyasu− Nagle/Strickland-Constable (NSC) soot model,24 even made significant contributions to broaden the understanding of soot behavior and, in hence, optimized the engine design and relevant control strategies to improve engine performance. With simplified chemical reactions, an empirical soot model is capable of easily implementing into computational fluid dynamics (CFD) programs to provide time-efficiency analysis. Therefore, it is essential and meaningful to develop an empirical soot model for biodiesel. The objective of the current study is to set up a phenomenological soot model of real biodiesel to explore the effects of combining the use of temperature control and biodiesel fuel on soot formation and oxidation mechanisms under diesel-like operation conditions. Experiments were conducted under 700, 800, 900, and 1000 K initial ambient temperatures, and chamber pressure, heat release rate, and time-related soot behavior were measured to validate the biodiesel soot model. After validation, the proposed biodiesel phenomenological soot model was applied to predicted transit behavior and two-dimensional distributions of soot-relevant intermediate species, such as acetylene, soot precursor, and OH radicals, under test conditions.
Figure 1. Schematic of the constant volume chamber. mounted, and the relevant configurations were tabulated as Table 1. To mimic the realistic diesel engine operation conditions, the cylinder
Table 1. Configurations of the HEUI 300A Injector and Fuel Injection Conditions parameter
value
nozzle style number of nozzle holes spray angle (deg) orifice diameter (mm) injection pressure (MPa) injection duration (ms) fuel volume (mm3) fuel temperature (K)
valve-covered orifice 6 140 0.145 134 3.5 120 350
wall was heated to 380 K by eight heaters made by Watlow Firerod, and the temperature of the oil and fuel lines inside the chamber head were kept at 350 K. The chamber pressure was measured by a quartz pressure transducer (Kistler 6121) embedded in the chamber wall in conjunction with a charge amplifier. As shown in Figure 1, images were obtained using a high-speed digital camera (Phantom, version 7.1) above the optical chamber and a light source was provided by a copper-vapor laser (Oxford Lasers LS20-50). Two-color outputs at 511 and 578 nm with a power ratio of 2:1 were provided by the copper-vapor laser. The high-speed camera and the copper-vapor laser were set to be synchronized up to 15 037 frames per second to provide time-related records at a resolution of 256 × 256 pixels. Only one of the six spray jets was examined in the experiment. A 105 mm focal length lens made by Nikkor with a maximum aperture of f4.5 was adopted for the images taking. To suppress the flame emission, two interference filters at 510 and 515 nm with 10 nm full width at half maximum (fwhm) were adopted to filter out the light at 578 nm. A 5 nm fwhm could be achieved when these two filters aligned together. Before laser light entered into the test constant volume chamber, it had been condensed by an condenser through a 6 mm diameter reflecting mirror in front of the condenser lens to be a form of point source. A high-speed camera was triggered by an injection signal and was set to record the whole combustion process of biodiesel fuel. 2.2. Experimental Methods. In this experiment, the soot diagnostic technology was a forward-illumination light-extinction (FILE) method developed by Xu and Lee,25 which was able to provide a two-dimensional time-resolved quantitative soot measurement. As shown in Figure 1, the light source and camera in the FILE method were placed at the same side of the flame through the same window. By doing this, only one window with a light diffuser was needed in the forward-illumination setup. In comparison to the backillumination light-extinction method, in which two aligned windows were required, the FILE technique was easier to operate. The light
2. EXPERIMENTAL SECTION 2.1. Experimental Apparatus. Experiments were conducted in an optical accessible constant volume chamber with a bore of 110 mm and a height of 65 mm. The chamber was designed to imitate spray and combustion processes in compressed ignition engines with a maximum operating pressure of 18 MPa. Figure 1 shows the schematic of the chamber together with the liquid spray scattering. To pass laser beams and take photographs, a fused silica (Dynasil 1100) was installed as an end window opposite the injector with 130 mm in diameter, 60 mm thick, and a high ultraviolet (UV) transmittance down to 190 nm. At the center of the chamber head, a hydraulicactuated electronic-controlled unit injector (HEUI, Caterpillar) was 5321
dx.doi.org/10.1021/ef401208b | Energy Fuels 2013, 27, 5320−5331
Energy & Fuels
Article
where ζ denotes the amount of excess oxygen, which is used to simulate an ambient oxygen concentration for diesel compressed ignition combustion. After acetylene was completely consumed, ambient air contains 21% oxygen, 66.7% nitrogen, 8.2% carbon dioxide, and 4.1% water vapor by volume. The density of the mixture for post-combustion is 14.8 kg/m3, which mimics the operation condition without an exhaust gas recirculation (EGR) rate in realistic diesel engines. The vessel pressure slowly decreases because of heat transfer through vessel walls. When the desired pressure was achieved, the injection signal triggered the HEUI injector and high-speed camera simultaneously, biodiesel fuel was injected into the cylinder, and the camera began to record the whole injection, autoignition, and combustion processes. In this paper, four different ambient temperatures were investigated, 700, 800, 900, and 1000 K, covering both low-temperature and conventional high-temperature combustion modes in diesel engines; an initial temperature of 700−800 K presents low-temperature combustion, and a 1000 K initial temperature presents typical hightemperature combustion. The test fuel in this experiment is soybean biodiesel, and its key properties were listed in Table 2.
diffuser set behind the flame was to ensure the sufficient reflected light that could be collected by the camera. As shown in Figure 2, the laser beam went through the soot cloud twice in the forward-illumination technique. Therefore, the light
Figure 2. Light extinction by the soot cloud.
intensity should be adjusted by the extinction because of the light diffuser and soot absorption. The variation of reflected light intensity was only caused by the presence of soot following Lambert−Beer’s law
I = I0 exp(−
∫0
Table 2. Properties of Soybean Biodiesel
2L
Kext dx)
densitya (g/cm3) viscosityb (mm2/s) cetane number sulfur (ppm) boiling point (K) lower heating value (MJ/kg) latent heat (kJ/kg) stoichiometric ratio oxygen content (%)
(1)
where I is the reflected light intensity with or without the presence of the soot cloud, I0 is the reflected light intensity without the presence of the soot cloud, Kext is the extinction coefficient, and L is the path length through the soot cloud. Unlike the traditional back-illumination method, light extinction in the FILE method is proportional to 2L rather than L because of the two passes through the soot cloud. The extinction coefficient of the soot cloud is dependent upon the particle number density, particle diameter, and optical properties. On the basis of Rayleigh approximation, the soot volume fraction can be expressed as
Cv =
⎛I ⎞ λ ln⎜ 0 ⎟ (2L)K a ⎝ I ⎠
a
3. DESCRIPTION OF SIMULATION MODELS 3.1. Phenomenological Biodiesel Soot Model. To further interpret the combining effects of the temperature and biodiesel on soot formation and oxidation mechanisms, a phenomenological soot model of real biodiesel was proposed to simulate soot behavior under various initial ambient temperatures (700, 800, 900, and 1000 K). The schematic of the phenomenological biodiesel soot model is presented in Figure 3. As shown in Figure 3, there are mainly nine steps in the model: (1) biodiesel pyrolysis process, (2) soot precursor formation, (3) soot inception reaction, (4) particle coagulation, (5) surface growth of soot nuclei, (6) soot surface oxidation via oxygen attachment, (7) soot surface oxidation via OH radicals, (8) acetylene oxidized by O2, and (9) precursor radicals oxidized by OH radicals. To improve the computation efficiency, all chemical reactions in this soot model are expressed in the form of a one-step global reaction. Detailed descriptions are given below. 3.1.1. Step 1: Biodiesel Pyrolysis. Unlike gasoline or diesel, which consists of hundreds of chemical compounds, biodiesel fuels only contain a limited number of compounds. A previous study confirmed that biodiesel derived from soybean and rapeseed oils (the most feedstock for biodiesel) is mainly composed of five long-chain C16 and C18 saturated and unsaturated methyl esters. The chemical structure and corresponding fractions of these five components in rapeseed methyl ester (RME) fuel and soy methyl ester (SME) fuel are listed in Table 3. Because C18 chain methyl esters are most abundant, methyl stearate (C18H36O2) was selected to be the
(3)
Equation 3 indicated that each pixel value represented local soot mass in each column vertical to the image plane, and the total soot mass at this time can be obtained by summing the pixel values all together. A detailed introduction about the FILE methods in detecting spray, combustion flame, and soot emissions can be found in previous studies.25,27−29 2.3. Experimental Procedure. In the beginning, a premixed, combustible gas mixture mixed by acetylene (C2H2), air, and nitrogen filled the test chamber. After the mixture was ignited by spark plugs, a high-temperature, high-pressure environment modeling a typical diesel in-cylinder environment with the piston at top dead center (TDC) was created in the test chamber. Because of flammability and low window contamination, acetylene, with unity C/H ratio, is adopted here as the pilot combustion gas. The density of the filling mixture can be determined by the chemical reaction as
4C2H 2 + (10 + ζ )O2 + 65N2 → 8CO2 + 4H 2O + ζ O2 + 65N2
At 25 °C. bAt 40 °C.
(2)
where λ is the wavelength of the monocolor light and Ka is the dimensionless absorption constant determined by soot refractive index m. A value of 5.47 is adopted here with m = 1.62 + i0.66. The soot volume fraction can be calculated using eq 2 through the analysis of images with and without soot clouds pixel by pixel. However, for a non-axisymmetric biodiesel flame, the thickness of the soot cloud cannot be measured and the only thing that is able to be detected is line of sight expressed as CvL. If the area of each pixel is represented by a dimension of Δr and the mean mass density of soot particles is adopted as 2.0 g/cm,26 then the soot mass at each pixel could be calculated as
mi = ρs CvLΔr
0.885 4.11 47.5
