1 AME 514 Applications of CombustionLecture 14: New Technologies

2 Emerging Technologies in Reacting Flows (Lecture 2)Applications of combustion (aka “chemically reacting flow”) knowledge to other fields (Lecture 1) Frontal polymerization Bacteria growth Inertial confinement fusion Astrophysical combustion New technologies (Lecture 2) Transient plasma ignition HCCI engines Combustion synthesis of photovoltaic cells Enhanced oil recovery using in situ combustion Microbial fuel cells Future needs in combustion research (Lecture 3) AME Spring Lecture 14

3 Transient plasma ignition - motivationMulti-point ignition of flames has potential to increase burning rates in many types of combustion engines, e.g. Pulse Detonation Engines Reciprocating Internal Combustion Engines High altitude restart of gas turbines Lasers, multi-point sparks challenging Lasers: energy efficiency, windows, fiber optics Multi-point sparks: multiple intrusive electrodes How to obtain multi-point, energy efficient ignition? Transient Plasma Ignition Flame spread over solid fuels is a useful model for simple 2 phase spreading flames, like building fires. It has well defined properties which can easily be quantified, like the spread rate which relates to CO production. And it is highly dependant on the fuel and environment. Downward, opposed flow flame spread at 1g is generally well understood. However, opposed flow flame spread at microgravity is less understood but important in manned spacecraft. upward, concurrent flow flame spread, which is especially important in a building fire, is also less understood. AME Spring Lecture 14 2

4 Transient plasma dischargesAlso called “pulsed corona” discharges Not to be confused with “plasma torch” Initial phase of spark discharge (< 100 ns) - highly conductive (arc) channel not yet formed Characteristics Multiple streamers of electrons High energy (10s of eV) electrons compared to sparks (~1 eV) Low anode & cathode drops, little radiation & shock formation - more efficient use of energy deposited into gas Flame spread over solid fuels is a useful model for simple 2 phase spreading flames, like building fires. It has well defined properties which can easily be quantified, like the spread rate which relates to CO production. And it is highly dependant on the fuel and environment. Downward, opposed flow flame spread at 1g is generally well understood. However, opposed flow flame spread at microgravity is less understood but important in manned spacecraft. upward, concurrent flow flame spread, which is especially important in a building fire, is also less understood. AME Spring Lecture 14 2

5 Characteristics of transient plasma dischargesFor short durations (1's to 100's of ns depending on pressure, geometry, gas, etc.) DC breakdown threshold of gas can be exceeded without breakdown if high voltage pulse can be created and stopped quickly enough AME Spring Lecture 14

6 Characteristics of transient plasma dischargesTransient plasma only Transient plasma + arc If arc forms, current increases some but voltage drops more, thus higher consumption of capacitor energy with little increase in energy deposited in gas (still have corona, but followed by (relatively ineffective) arc) AME Spring Lecture 14

7 Transient plamsa discharges are energy-efficientDischarge efficiency d ≈ 10x higher for transient plasma than for conventional sparks AME Spring Lecture 14

8 Engines 101 - slow burn reduces efficiencyBurn starts earlier in compression process, has to go to same v, result is higher s to get same heat addition (= ∫ Tds) Difference in work: 2 triangular slivers vs. rectangle Burning before or after piston/cylinder reaches its minimum volume ALWAYS leads to lower efficiency since it leads to lower TH for same TL, thus lower efficiency Carnot “strips” AME Spring Lecture 14

9 Engine experiments Theiss et al., 20072000 Ford Ranger I-4 engine with dual-plug head to test transient plasma & spark at same time, same operating conditions National Instruments / Labview data acquisition & control Horiba emissions bench Pressure / volume measurements Optical Encoder mounted to crankshaft Spark plug mounted Kistler piezoelectric pressure transducer AME Spring Lecture 14

10 Electrode configurationMacor machinable ceramic insulator Coaxial shielded cable “Point to plane” geometry - by no means optimal AME Spring Lecture 14

11 On-engine results Transient-plasma (corona) ignition shows increase in peak pressure under all conditions tested Cylinder pressure (pounds/in2) AME Spring Lecture 14

12 On-engine results Transient plasma ignition shows increase in Indicated Mean Effective Pressure (IMEP) under all conditions tested Cylinder pressure (pounds/in2) Vmax Vmin AME Spring Lecture 14

13 IMEP at various air / fuel ratiosIMEP higher for transient plasma than spark, especially for lean mixtures (nearly 30%) Coefficient of variance (COV) comparable AME Spring Lecture 14

14 Burn rates Integrated heat release shows faster burning with transient plasma leads to greater effective heat release 2900 RPM,  = 0.7 AME Spring Lecture 14

15 Burn rates Transient plasma ignition shows substantially faster burn rates at same conditions compared to 2-plug conventional ignition 2900 RPM,  = 0.7 AME Spring Lecture 14

16 Emissions data - NOx Improved NOx performance vs. indicated efficiency tradeoff compared to spark ignition by using leaner mixtures with sufficiently rapid burning (CO, UHC similar to spark ignition), though improvement isn't as impressive as one might hope because in general faster burning means high peak T, thus higher NO – but maybe there's a more intelligent way to use the technology… AME Spring Lecture 14

17 Heat transfer in engines - catechismWhy do we have heat loss in IC engines? Because the cylinder wall is “cold” - typically just a little higher than the cooling water temperature, ≈ 120˚C (boiling point at 2 atm). This is much colder than the gases during combustion (2400K) and during expansion (down to 1200K). Why do we need to cool the cylinder? To keep the lubricating oil from getting too hot and breaking down. Also, with too large a temperature increase, thermal expansion will change the fit between the piston and cylinder and make it too tight or too loose. How significant is the loss? See Heywood (1988, Fig. 12-4): At low vehicle speed (meaning: low engine RPM, low Pintake) 50% of fuel energy is dissipated as cooling system losses; at higher speed, 30%; Heywood (p. 851) states that a 10% decrease in heat loss would mean about 3% increase efficiency Could we reduce the loss by using a ceramic (or whatever material) engine that could withstand high temperatures without oil lubrication? Analysis shows that raising the wall temperature doesn’t help much, what is needed is a more nearly adiabatic engine (lower heat transfer coefficient). This is borne out by many experiments, peaking in the 1980's, using so-called "adiabatic" engines made of ceramics. This raised the wall temperature, but then you had more heat transfer during compression, thereby increasing compression work, so the efficiency didn’t improve. AME Spring Lecture 14

18 Heat transfer in engines - catechismHow can we decrease the heat transfer coefficient (h)? Heat transfer in engines is controlled by turbulence, so you'd need to decrease turbulence How can you do that? Engines are normally designed for high turbulence, so you could reverse-engineer the engine for lower turbulence (e.g. by avoiding swirl in the intake ports) Why don't we do that now? Because we need high turbulence (high u') to get fast burning Is there any way to burn fast without turbulence? Well DUH, TPI as we've just discussed Doesn't lower turbulence have a negative impact? Maybe, but actually it has at least one additional solid benefit; decreasing heat loss means that a smaller radiator can be used; the weight and cost of the radiator wouldn't be much less but the aerodynamic drag caused by radiators is significant and reducing the radiator frontal area could cause a noticeable increase in fuel economy just by reducing drag. Also there would be a slight increase in volumetric efficiency (the amount of air that can get into the engine; more air  more fuel can be burned  more heat release  more power) because you're not forcing the air to follow a tumultuous path in order to get enough swirl to get high turbulence; with a more direct path there would be less pressure drop in the intake system and thus higher intake air density AME Spring Lecture 14

19 Transient plasma ignition for PDEsSignificant limitation of Pulsed Detonation Engines is Deflagration to Detonation Transition (DDT) distance - if too long, can't get DDT within engine! Formation of initial pressure waves from deflagration is typically slowest step & most easily accelerated by transient plasma ignition Pre-DDT region has lower pressure Detonation tube experiments (Lieberman et al., 2005) Stoichiometric C2H4-O2 with N2 dilution (air = 73.9% dilution) Spark plug (SP) vs Transient Plasma (PI) AME Spring Lecture 14

20 Transient plasma ignition for PDEsMuch lower ignition delay times (time to first recorded pressure rise at Pcb1) with PI, with or without turbulence-generating obstacles (spiral insert) (accelerates DDT but causes heat & pressure losses), so can have higher repetition rates, thus more thrust … but PI doesn’t effect Specific Impulse (ISP) nearly as much In these plots ISP is defined as (Impulse / weight of mixture) not (Impulse / weight of fuel), so ISP values are very low AME Spring Lecture 14

21 HCCI engines Burning rapidly at minimum volume yields the best possible thermal efficiency, but damage due to knocking means we want to burn “fast but not too fast” HCCI - Homogeneous Charge Compression Ignition engines take advantages of this - “controlled knocking” By using homogeneous reaction instead of flame propagation, conventional flammability/misfire limits absent Can burn very lean mixtures, low Tad, low peak temperature, low NOx formation Lean mixtures - can obtain part-load operation without throttling and its losses Since we're asking for knock, use high compression ratios, thus high th AME Spring Lecture 14

22 HCCI engines Standard engine Knocking engine HCCI engineVideos courtesy Prof. Yuji Ikeda, Kobe University AME Spring Lecture 14

23 Comparison of gasoline, diesel & HCCIAME Spring Lecture 14

24 Comparison of gasoline, diesel & HCCIJ. Dec., Proc. Combust. Inst., Vol. 32, p (2009). AME Spring Lecture 14

25 HCCI engines Much more difficult to control the rate and timing of homogenous reaction than a propagating spark-ignited flame; various control schemes being studied Variable intake temperature Variable exhaust gas recirculation Variable compression ratio and valve timing Cycle-to-cycle control probably needed HCCI experiments in a single-cylinder engine AME Spring Lecture 14

26 HCCI experiments in 6 cyl. engine (1 cyl. HCCI)Dec and Sjöberg, 2002 AME Spring Lecture 14

27 HCCI control using mixture ratioShaver et al., 2009 Control peak pressure (minimize engine noise) using closed-loop mixture ratio control AME Spring Lecture 14

28 HCCI - disadvantages (opportunities?)Difficult to control timing and rate of combustion If misfire occurs, gas mixture during the next cycle will be too cold for auto-ignition to occur (unless intake air heating is used), the engine will stop Cold starting? Operating window for HCCI operation (load and engine speed) is small - most HCCI concepts use conventional spark-ignited operation at higher loads (less lean mixtures) Additional components for control system - increased cost Relatively high friction losses due to low IMEP, thus friction loss is a higher % of net work (indicated work - friction) AME Spring Lecture 14

29 Combustion synthesis of materials for PV cellsCourtesy of Prof. Hai Wang, Stanford/USC (Tolmachoff et al., 2009) Current photovoltaic (solar) cells are reasonably efficient but very expensive to produce (≈ $10/watt vs. $1/watt for conventional electric power); net cost of solar ≈ 5x conventional power Dye-sensitized solar cells not as efficient but cheap to manufacture First proposed by O’Regan and Grätzel (1991) Somewhat like fuel cell Anode: transparent, conductive glass, coated with TiO2 nanoparticles in turn coated with fluorescent dye to absorb incoming photons Electrolyte: I- / I3- oxidation/reduction reaction – basically a diode so current can only flow one direction Cathode: Pt-coated transparent, conductive glass AME Spring Lecture 14

30 Dye-Sensitized Solar CellTransparent conducting glass Transparent conducting glass TiO2 dye electrolyte -0.5 0.0 0.5 1.0 E (V) e- S* hu maximum Voltage ~0.75 V ox (I3-) red (I-) Redox mediator Tradeoff: small particles have high surface area, so pick up more photons; large partlcies have fewer necks, so transport electrons less resistance – optimal exists. TiO2 has a work function such that once an electron is in a conduction band it stays there and cannot jump to a valence band. (Unlike silicon). As long as the particle is truly crystaline, the electron won’t fall. (Performance is limited by regularlity of the crystal. I3- + 2e-  3 I- S e- AME Spring Lecture 14

31 TiO2 particle considerationsTiO2 has advantages over silicon - TiO2 “work function” such that once an electron jumps to conduction band it stays; cannot fall back down to valence band (if particle truly crystalline) Ideal particle size < 10 nm Too large: low surface/volume ratio, don’t get good electron collection Too small: too many contacts between particles, causes high resistance to electron flow Current technique for anode fabrication Commercial TiO2 powder (> 20 nm) Making a paste/paint & screen printing Sinter at 450◦C (glass substrate only) Stain with dye Wang method Particle synthesis and film deposition in one step No need to sinter AME Spring Lecture 14

32 Synthesis method – stagnation flameFlame Stabilizer Tmax Stagnation flame vO burner-stabilized flame Tubular burner Carrier gas Ar vO Shielding Ar C2H4/O2/Ar TTIP/Ar Particle properties controlled by Flame temperature (argon dilution) Reaction time (flow rate) Ti precursor concentration TTIP, (Titanium Isopropoxide) TTIP Electric mantle AME Spring Lecture 14

33 Flame Structure (C2H4-O2-Ar, f = 0.4)500 1000 1500 2000 2500 Stagnation surface (x = 3.4 cm) T (K) Particle nucleation/ growth region 100 200 300 400 500 Axial Velocity v (cm/s) Laminar flame speed Particle nucleation/ growth region Growth time limited to 2 ms because of thermophoresis. On increasing T side of flame, convection is rapid and TP can’t hold particles, but as particle approaches stagnation plane, U decreases and TP force pushes particle along faster (about 1 m/s), limiting growth time and thus particle size. Also very uniform residence time for on-axis and off-axis particles. 10 H O CO O 2 2 2 10 -1 C H 2 4 CO Mole Fraction 10 -2 10 -3 H 2 H Computations using Sandia counterflow flame code and USC Mech II 10 -4 2.7 2.8 2.9 3.0 3.1 3.2 3.3 Distance from nozzle, x (cm) AME Spring Lecture 14

34 Aspects of particle growthGrowth time limited to 2 ms because of thermophoresis (TP) – moves particles to from high T to low T in gas; for particles much smaller than mean free path (Good website on thermophoresis: On increasing T side of flame, convection is rapid and TP can’t hold particles in place against convection As particle approaches stagnation plane, V decreases and TP force pushes particle along faster (≈ SL), limiting growth time and thus particle size Very uniform residence time for on-axis and off-axis particles – characteristic of stagnation flow AME Spring Lecture 14

35 Particle size distributionsIncrease Ti Precursor Concentration Particle size can be well controlled Size distribution widens as median size increases but the size variation still small compared to other methods AME Spring Lecture 14

36 Flame Stabilized on Rotating SurfaceWant boundary layer thickness d ~ (n/wrad)1/2 < distance from flame to stagnation surface, so rotation doesn’t affect particle formation & growth ~0.3 cm AME Spring Lecture 14

37 Stationary vs. Rotating Stagnation PlateRotating the stagnation surface results in smaller particles and narrower distributions AME Spring Lecture 14

38 Comparison with commercial TiO2Tested under the standard AM1.5 solar light Use Solaronix purple dye, comparisons made under comparable conditions FSTS films (largely unoptimized) outperform Degussa powder with screening printing Method allows continuous, reel-to-reel fabrication of DSSC photoanodes in one step AME Spring Lecture 14

39 “In situ” enhanced oil recoverySee review by Mahinpey (2007) Heavy-oil reservoirs containing high-viscosity oil are impossible to produce via conventional pumping Viscosity decrease via steam injection expensive & of limited effectiveness Can inject air and combust a portion of oil Has seen limited field use but can be effective Limited laboratory experiments, in small diameter tubes Real situation: large cross-section – instabilities Similar to “filtration combustion” of porous solid Very similar to flames in Hele-Shaw cell (see lecture 8) Flow described by Darcy’s Law Buoyancy (RT), thermal expansion (DL), viscosity change (ST) instabilities …. But a non-premixed, 3-phase (air, oil, inert porous solid) system! Practical limitation – emissions! How to conduct laboratory experiments that are relevant to oil field production? AME Spring Lecture 14

40 “In situ” enhanced oil recovery AME Spring Lecture 14

41 “In situ” enhanced oil recoveryKök, et al. (2008) AME Spring Lecture 14

42 References AME 514 - Spring 2017 - Lecture 14Dec, J. E., Sjöberg, M. (2002). “HCCI Combustion: The Sources of Emissions at Low Loads and the Effects of GDI Fuel Injection.” 2002 Diesel Engine Emissions Reductions (DEER) Conference, San Diego, CA, August (http://energy.gov/sites/prod/files/2014/03/f9/2002_deer_dec.pdf) Heywood, J. B., “Internal Combustion Engine Fundamentals,” McGraw-Hill, 1988 M. V. Kök, G. Guner, S. Bagci, Oil Shale, Vol. 25, No. 2, pp. 217–225 (2008) D. Lieberman, J. Shepherd, F. Wang and M. Gundersen (2005). “Characterization of a Corona Discharge Initiator Using Detonation Tube Impulse Measurements,” 43rd AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan , 2005 (AIAA Paper ). Mahinpey, N., Ambalae, A., Asghari,K., “In situ combustion in enhanced oil recovery (EOR): A review,” Chem. Eng. Commun. Vol. 194, pp (2007) B. O’Regan and M. Grätzel, Nature 343, , 1991. Shaver, G. M., Gerdes, J. C., Roelle, M. J. (2009). “Physics-Based Modeling and Control of Residual-Affected HCCI Engines,” J. Dyn. Sys., Meas., Control 131(2), N. Theiss, J. B. Liu, J. Levin, J. Zhao, T. Shimizu, F. Wang, M. A. Gundersen, P. D. Ronney (2007). “Transient Plasma Discharge Ignition for Internal Combustion Engines,” 2007 U. S. National Meeting, Combustion Institute, La Jolla, CA, March 26 – 28, 2007. E. D. Tolmachoff, A. D. Abid, D. J. Phares, C. S. Campbell, H. Wang (2009). “Synthesis of nano-phase TiO2 crystalline films over premixed stagnation flames,” Proc. Combust. Inst. 32, pp. 1839–1845. Wang, F., Liu, J. B., Sinibaldi, J., Brophy, C., Kuthi, A., Jiang, C., Ronney, P. D., Gundersen, M. A. (2005). "Transient Plasma Ignition of Quiescent and Flowing Fuel Mixtures, " IEEE Transactions on Plasma Science, 33: AME Spring Lecture 14