Adiabatic Point Research Articles

Advanced coolside desulfurization process

AbstractThis paper describes a dry sorbent injection SO2 abatement process with the potential to significantly enhance desulfurization performance over that of existing in‐duct injection technology such as the Coolside process. The Advanced Coolside process was developed by CONSOL Inc. and is being tested in a pilotscale project, partially funded by the U.S. Department of Energy (DOE). The goals of this process development effort are to increase the applicability of sorbent injection technology as a compliance option for coal‐fired boilers for the 1990 Clean Air Act Amendments and to reduce total process SO2 control costs. Specific performance objectives are to achieve 90% SO2 removal and 60% sorbent utilization efficiency, while retaining the low‐capital cost and retrofit advantages inherent to in‐duct sorbent injection technology. The Advanced Coolside process concept involves flue gas humidification to near the adiabatic saturation point combined with fly ash removal using a gas/liquid contacting device, followed by injection of a sorbent (e.g., Ca(OH)2) into the highly humid flue gas with collection by the existing particulate collector. High SO2 removal and sorbent utilization efficiency are possible with hydrated lime when the flue gas is nearly saturated with moisture. Sorbent utilization is further improved by recycle, and the potential for improvement by the use of recycle is increased with removal of the fly ash before sorbent injection. Further, testing indicates that pre‐treatment of the recycle sorbent before reinjection can improve process performance. Because flue gas is humidified completely before sorbent injection, the existing duct configuration may not limit process retrofit potential. The heat of reaction between SO2 and Ca(OH)2 heats the flue gas before the existing particulate collector. Current pilot test data indicate that the Advanced Coolside process performance objectives are achievable. On‐going process development work includes design optimization for the contacting device, improving efficiency of sorbent recycle, and optimization of sorbent performance.

Characterization of flame spread over PMMA using holographic interferometry sample orientation effects

Flame spread over a thermally thick slab of PMMA at several angles of sample orientation from θ = −90° (vertically downward flame spread) to θ = +90° (vertically upward flame spread) in air was investigated by measuring temperature distributions within the PMMA sample in the vicinity of the leading edge of the flame front using holographic interferometry. Samples with widths of 0.32, 0.47, 1.0, and 2.5 cm, a thickness of 2.5 cm, and a length of 30 cm were used. The measured net heat flux from the gas phase to the sample surface at the vaporization front of the sample is about 7 W/cm 2 for downward flame spread (θ < 0δ), 6.5 W/cm 2 at θ = + 10°, and 2.8 W/cm 2 at θ = + 90°. However, the total net heat transfer rate increases with an increase in the angle of sample orientation, because the characteristic heating length, defined as the distance from the adiabatic point on the sample surface to the vaporization point, increases with an increase in the orientation angle of the sample. The total net heat transfer rate into the sample from the gas phase is about 56% of the total net heat transfer input to the sample at θ = − 90°, 78% at θ = 0°, 87% at θ = + 10°, and 99% at θ = + 90°. Therefore, heat transfer from the gas phase into the unburnt fuel ahead of the vaporization point is the dominant heat tranfer path for all angles of orientation. This was clearly demonstrated by the net heat flow vector patterns within the sample. The streamwise conductive heat transfer rate through the sample decreases with an increase in flame spread rate (increase in the sample orientation angle) due to insufficient time being available for the slow thermal wave to travel through the sample.

Limestone based spray dry scrubbing of SO 2

Spray dry scrubbing for emissions control has emerged during recent years. Worldwide, more than fifty spray dry scrubbers are in operation or under construction for the control of SO 2 on coal fired boilers. Virtually all of these units utilize a slurry of slaked lime as a reactive agent for SO 2 capture. The cost of lime is several times the cost of limestone, so it would be a great advantage if limestone could be used in the process. Research on spray dry scrubbing has been underway for over six years at our department. The investigation on which the present paper is based was undertaken to study the prospects for spray dry scrubbing of SO 2 by utilizing a slurry of finely ground limestone. The experiments were performed in a 0.5 MW th pilot plant, consisting of a spray dryer and a fabric filter system. The slurry was prepared from limestone ground to various particle sizes. Parameters investigated were the approach to the adiabatic saturation point, the stoichiometric ratio (CaCO 3/SO 2), and the addition of CaCl 2, which would result from the chlorine in the coal. Furthermore, the impact of the pressure drop over the fabric filter on SO 2 removal was studied. The experimental results showed that ground limestone can remove SO 2 efficiently in a spray dry scrubber system. A significant contribution to the removal occurs in the fabric filter. The removal increases as the surface area of the limestone increases due to grinding. However, a limiting value of the removal over the spray dryer is obtained at quite modest surface areas. This is caused by agglomeration of smaller particles at low pH values; as the droplets are contacted with the flue gas, the pH drops due to hydrogen ion production from hydrolysis of SO 2. This was seen by measuring the paricle size distribution at different pH values. Hence, while the gas—solid reaction in the fabric filter depends on the surface area, the rate for the gas—liquid—solid reaction in the droplets depends on the particle or agglomerate size.

On the influence of the plate thickness on the boundary layer ignition for large activation energies

An analysis is performed of the influence of the plate thermal resistance on the boundary layer thermal ignition of a premixed combustible gas for the case of large activation energy. The analysis makes use of the boundary layer approximation to describe the gas flow and of first-order matched asymptotic expansions to define ignition. Explicit expressions are derived for the critical Damköhler number for ignition and for the ignition distance. It is shown that these variables depend on a nondimensional parameter that includes the plate and gas thermal resistances (Biot number) and the local Damköhler number. The magnitude of the ignition variables increases with this parameter. Comparison of the results for high and low activation energies shows that their magnitude affects not only the location of ignition but the relative position of the adiabatic and ignition points.

Numerical solution of a nonsteady blast wave propagation in two-phase (“separated flow”) reactive medium

The problem of a detonation in a two-phase reactive medium is fully solved using the flux corrected transport (FCT) algorithm. This method is used to solve nonlinear hyperbolic equations of two-phase flow and was chosen for its ability to suppress numerical oscillations and to describe well the complex shock waves. The solution of the problem is described. The details of a test case based on the analytical solution of a strong adiabatic point explosion are given.

Transient phenomena in boundary layer ignition with finite plate thermal resistance

In this paper the transient ignition of a gaseous combustible mixture in a flat plate boundary layer is studied. The thermal capacitance of the plate is neglected but its thermal resistance is taken into account. The lower surface of the plate is given a sudden increase in temperature and the processes leading to a permanent flame structure are analyzed through numerical solutions of the parabolic governing equations. The events leading to ignition are characterized by the formation and propagation downstream of a high reaction rate spot. The generation of this spot is related to the appearance of a zero heat transfer point on the wall. Three characteristic times can then be defined for the ignition process: (a) the time necessary for the formation of the adiabatic point on the wall, (b) the time necessary for the high reaction rate spot to reach the steady state flame position, and (c) the time after which a removal of the temperature perturbation at the plate lower surface would still lead to flame formation within a reasonable distance. These characteristic times are found to be strongly affected by the plate thermal resistance.

Blast waves in real gases

Real gas effects of compressibility, vibrational excitation, dissociation, ionization and electronic excitation are taken into account in blast wave formulation. Residual property concepts, as well as those of equilibrium quantum statistical thermodynamics, are utilized to relate the different microscopic internal energy excitation levels to macroscopic thermodynamic properties. Jump conditions across the shock front are then evaluated and a non-self-similar quasi-similar solution is obtained for the adiabatic point explosion problem. From the results it appears that temperature profiles are affected most by the real gas properties of the medium. Some of their effects are evident in density profiles, while pressure and particle velocity are relatively unaffected. Finally, the shock front decays slower than in the perfect gas case, especially during the initial stages of propagation.

Collisional ionization and elastic scattering in alkali–halogen atom collisions

Close-coupling (two-state) computations for collisional ionization and elastic scattering in Na+I and Li+I collisions have been made, using realistic model potentials, at a large range of postthreshold energies. At the higher energies a generalized JWKB procedure was employed which yields a very accurate scattering matrix when compared to a fully quantal solution. The computed elastic and inelastic total and differential cross sections are discussed and compared to the predictions of more approximate methods. The rainbow angles in the angular distribution are compared to those predicted by the classical deflection function obtained from the trajectories along adiabatic or diabatic potential curves (based on the assumption of a single transition point). It is found that at all energies considered these rainbow positions agree best for trajectories computed via the adiabatic representation rather than the diabatic representation. The double rainbow pattern noted in the elastic ionic channel can be interpreted only within the adiabatic representation. Landau–Zener–Stueckelberg (LZS) computations for the energy dependence of the collisional ionization cross section have also been made. The LZS curve jumping mechanism is in better agreement with the quantal results when the transition is taken to occur at the adiabatic coupling point (Rp) rather than the diabatic crossing point (Rx). Moreover, depending on the nature of the potentials, there may also be an additional transition region at small internuclear distances, which is not present in the LZX scheme but may be the source of oscillations found in the post maximum energy dependence of the collisional ionization cross section. The systematics of the collisional ionization cross sections for all alkali metal–halogen atom (M+X) collisions are discussed and the energy dependence of the cross section is computed using realistic potentials. In terms of the reduced velocity (at the crossing point) v/Kp, it is noted that only Li+I and Na+I are in the strong coupling region. For all other systems the cross section is found to decrease with increasing collision energy.

Dynamics of an Adiabatic Point Reactor

The problem of adiabatic excursions in a reactor is studied in general. We let the prompt temperature reactivity feedback be an unspecified function of temperature, ρ = ρ0 = ρ0 + f(T), where ρ is total reactivity, ρ0 initial step reactivity and f(T) the feedback function. The similarity of the behavior of the reactor for different f(T) is established by means of a topological (qualitative) analysis. A quantitative asymptotic solution of the non-linear system of DE describing the reactor is presented. In delayed critical excursions, the delayed neutrons play a determining role. In the first part of a prompt excursion, the delayed-neutron source is nil; however this is not so in the second part, where it contributes appreciably to the excursion. These conclusions are shown to be valid in general, and allow us to write down almost directly the (approximate) quantitative solution of the non-linear system for any f(T). These results are correlated with the experimental data for the adiabatic excursions of a UO2 core in SPERT I; in this case the (prompt) dependence of the reactivity on energy is of the form ρ = ρ0 - 4.588 × 10-4E0.74.

Kinetics and Mechanism of Ethylene Oxide Decomposition at High Temperatures

In a new flow reactor designed for the study of reaction kinetics at high temperatures, the rates of decomposition of ethylene oxide were measured to temperatures approaching its adiabatic decomposition point (∼1200°K). This device is unique in that it follows the reaction process, not by a freezing technique usually employed in flow systems, but by tracing the temperature history of the reaction. The ethylene oxide decomposition study showed that the kinetics followed an over-all first-order mechanism and the over-all activation energy was 42±2 kcal/mole. The results were found to be independent of the amount of fuel injected, the velocity of the inert carrier gas, and the inert gas itself. A free radical decomposition mechanism consistent with the experimental results is postulated.