All co-firing experiments were conducted using a 30 kW (100,000 BTU/h, approximately 15 lb or 6.80 kg of coal/h) small-scale furnace capable of firing most types of ground fuels. A schematic of the furnace is shown in Figure 3.20. Propane and natural gas are used to heat the furnace to the operating temperature of 1100°C (2000°F). Type K (shielded, ungrounded) thermocouples are used to measure the temperature along the axial length of the furnace. A solid fuel hopper feeds coal and coal/biomass blends during experiments. Primary air (6m3/h, 15-20% of total air) is necessary to propel the finely ground solid fuel through the fuel line and to the furnace. Prior to ventilation, all exhaust gases pass through a water-cooling spray to significantly lower the temperature of the gases. A sump pump pumps this water out of the furnace. More details are provided in Frazitta et al. (1999), Arumugam et al. (2006c), Annamalai et al. (2005), Lawrence et al. (2009) and in Thien et al. (2012).

The secondary air (75-80% of total air) heater was run for an hour before the experiment was started. The secondary air is swirled (Swirl number = 0.7) prior to entry into the combustion chamber. Once the secondary air reached a steady temperature, approximately 500 K (440.33°F), the propane torches were ignited. Natural gas and propane were used to preheat the furnace to operating temperatures. Once the furnace reached 1366 K (2000°F), the natural gas was turned off and the natural gas line was closed. The solid feeder line was opened and the solid feeder was turned on and set to the desired fuel flow rate. The primary and secondary air lines were set to the appropriate flow rates to obtain the desired equivalence ratio. The furnace was allowed to run for 30 minutes before the first readings were taken. The measurement was taken at the last sampling port just before the quenching water sprays and the wet flue gases were ducted to the atmosphere. After taking a measurement at this equivalence ratio, the secondary air could be adjusted to a different equivalence ratio. After taking measurements at all desired equivalence ratios, the furnace was turned off.

Fuel properties played a significant impact on the burnt fraction and the emissions created by combustion. The results from the co-firing experiments performed are discussed and their role in evaluating the combustion performance of the fuels is explained. The performance was evaluated by measuring combustion efficiency (burnt fraction) and the emissions levels of pollutants that include NOX and CO. In addition, overall fuel nitrogen conversion efficiency to NOX was also determined. The mercury emissions are presented elsewhere (Udayasarathy, 2007).

The co-firing involves a mix of finely pulverized biomass and coal. The size distribution was obtained using an ASTM sieve shaker. Sauter mean diameter (commonly abbreviated as SMD or d32) is commonly used for estimating the average size of solid fuel particles. The SMD is defined as the diameter of a sphere that has the same ratio of volume to surface area. It is represented as the following equation:

n

J2dl ■ n

SMD or d32 = -=————-

^ di ■ ni

i=1

where di is the diameter of particles and ni is the number of the particles of diameter di. According to the Rosin Rammler fit, the cumulative mass fraction CMF or drops (or particles) with a dimension lesser than dp is given as (Annamalai and Puri, 2007):

CMF = 1 — exp(-bdp)

where b: size constant, n: distribution constant and is a measure of spread of drop size. In terms

of the dp, charac size:

CMF = 1 — exp

where dp, charac denotes the characteristic drop or particle size for which CMF = 1 — exp(-1) = 63.2% and

1

dn

p, charac

The fraction R having size greater than dp is:

The plot of ln{R} vs. dp must be linear and the slope yields “n” and dp, charac is determined from the plot at R = 0.368. The values of “n”, “b” and SMD are presented in Table 3.9 for several fuels. Note that the coals had a larger SMD than those of DB fuels. The dirt (or mineral matter) that got collected with the DB fuels passed through all of the sieves and collected in the pan. This caused the DBs to have a smaller SMD.

Table 3.9. Size distribution parameters, adopted from Lawrence (2007).

Size distribution parameters

TXL WYO LA-PC-DB-SepS HA-PC-DB-SoilS

n 1.2991 1.4369 1.0934 1.2612

b 0.000934 0.00042 0.0024 0.0013

SMD (microns) 396 396 96.7 91.6

1.2.2 Introduction

Gasification is a technology commonly used nowadays for extracting energy from biomass. Over the past decade, there has been great progress in the development of gasification technology in China. Many kinds of biomass gasification processes have been developed, treating different materials for various purposes.

1.2.3 Gasification technology development

Dating back to the 19th century, gasification technology now attracts new interests in Europe, because of the end use flexibility of the syngas (Lan et al., 2011).

A charcoal gasifier had been developed in the 1940s and was tried to drive vehicles with the technology, which is the initial exploration in China in gasification of biomass. But the technology did not obtain further development for various reasons.

A fluidized bed reactor for industrial applications had been developed in China in the 1950s, but there were some imperfections in the technology, and the application was suspended. In the 1960s, Chinese researchers began to study the biomass gasification power generation and gained some experiences, and the preliminary prototype had been developed and gained some experiences, but these researches were stopped because of the economic conditions and the small profits.

A fixed bed gasifier (updraft and downdraft) circulating fluidized bed gasifer was developed in China in the 1980s. The product gases were used for power generation, supplying heat and cooking.

An 1 MW BGPG system with a circulating fluidized bed (CFB) gasifier had been developed, and constructed in a rice mill in the Fujian province of China (Wu et al., 2002; Wang et al., 2005; Lu et al., 2004b; Wu et al., 2003). A neural network was focused on for the simulation of gasification process. An artificial neural network model was developed to simulate the gasification processes in order to obtain the gasification profiles (Guo et al., 2001; Tang et al., 2003; Wang et al., 2002).

Gasification and polygeneration technology in the fluidized bed were concentrated on by Tie et al. (2003; 2005). They studied four kinds of biomass (bagasse, pine sawdust, peanut shell, rice husk) in a fluidized bed reactor and found that the temperature was a key parameter due to biomass pyrolysis in a fluidized bed reactor.

Lu etal. (2005; 2007) developed hydrogen production technology by using supercritical water. He and other researchers (Guo et al., 2006; Yan et al., 2006) mixed several kinds of biomass in sodium carboxymethylcellulose, which was gasified successfully at 650°C, 25 MPa in a tubular flow reactor with formation of hydrogen, carbon dioxide, carbon monoxide, methane and a small amount of ethane and ethylene.

Furthermore, the domestic garbage gasification was studied by Yuan et al. (2002). Chen et al. (2003a, b; 2005; 2006; 2008) made progress in cogasification of biomass and glycerin. They studied four kinds of biomass in a two-stage reactor to produce hydrogen-rich gas, and investigated the effect of a catalytic bed on the pyrolysis behavior.

In many chemical looping processes calcium sorbents are used to separate CO2 from the other gases. These reactions generally involve absorption/calcination processes, based on the following

CO? and SO}

Figure 5.28. The Alstom process (Rizeq et al., 2002; Andrus et al., 2006).

chemical reactions:

CaO + CO2 ^ CaCO3

+heat (5.36)

CaCO3 —CaO + CO2

These reactions can be shifted to the left or to the right by changing the temperature and pressure.

The combustion of carbonaceous fuels can be assumed as follows:

(lx + y/2 — z) MO + CxHyOz ^ (lx + y/2 — z)M + xCO2 + y/2H2O

M + Air ^ MO + N2 + unreacted O2 (5.37)

where the looping medium is a metal oxide. The first reaction only produces CO2 and steam; therefore carbon dioxide may be easily separated from steam by condensing the latter. The second reactor oxidizes the metal again and the flue gases are nitrogen and oxygen. The second reaction provides heat to the first reactor. The overall combustion process is thus divided into two sub processes which separate inherently the flue gas components.

If the fuel consists of carbon only, then the chemical looping reactions can be as follows:

2MO + C ^ 2M + CO2 M + H2O ^ MO + H2

The looping medium is the same as in the previous reactions, but the flue gas of the second reactors is now pure hydrogen. The process allows separating CO2 for capture and producing hydrogen as a byproduct.

The second reactor produces hydrogen instead of heat and steam is the oxidant of the overall combustion process. The carbon capture is greatly simplified and requires much less energy than in any other process producing hydrogen.

Hydrogen can also be produced including a calcination process in a chemical looping combustion:

CO (g) + H2O (g) ^ CO2 (g) + H2 (g) CO2 (g) + CaO (s) ^ CaCO3 (s) CaCOs (s) ^ CO2 (g) + CaO (s)

One positive effect of capturing CO2 in the process by means of CaO, is to enhance the hydrogen production from the first reaction because the chemical equilibrium favors its formation. The calciner in the second reactor allows an efficient separation of CO2 from the flue gas which can be ready for capture.

There has been relatively little work concerning high pressure syngas flames. McLean et al. (1994) and Vagelopoulos and Egolfopoulos (1998) reported premixed flame speeds at pressures from atmospheric to a few atmospheres. Burke et al. (2007) examined the effect of CO2 on burning velocity of spherically expanding flames at p = 1.0 and 10 atm using a 25%H2-75%CO mixture with 12.5%O2-87.5%He oxidizer. Sun et al. (2005) reported laminar flame speeds for

CO/H2/air and CO/H2/O2/He mixtures for pressures up to 40 atmospheres using the constant- pressure spherical flame technique. A kinetic model was also developed using the latest available thermo-transport and kinetic data (Park et al., 2004; Ouimette, 2009). The mechanism was validated against the measured flame speeds, non-premixed counter flow ignition temperatures, concentration profiles in a flow reactor, and ignition data from shock tube experiments. Figure 2.16 from their study shows the measured and predicted laminar flame speeds plotted versus Ф for CO/H2/O2/He mixtures at different CO/H2 ratios, and pressures of 5-40 atm.

Predictions are based on their kinetic model and that reported by Davis et al. (2005). As expected, the flame speed increases with increasing H2 content, and decreases with increasing pressure. Overall, there is good agreement between the predictions and measurements, although both models exhibit discrepancies, whichmay be attributed to uncertainties in kinetic and transport data. Thus, further studies are warranted for high-pressure syngas flames over a range of com­bustion regimes, including non-premixed and partially premixed combustion, and using different burners.

Studies on turbulent syngas flames have focused on the determination of turbulent flame speeds (ST) (Chase et al., 1951; Kee et al., 1995; Daniele, 2011). While ST can be defined in multiple ways, it is often based on a global consumption speed (Venkateswaran et al., 2011) and is presented in terms of the normalized flame speed (ST/SL) as a function of turbulence intensity, fuel composition and other parameters. Daniele et al. (2011) considered the reaction zones regime and examined the effects of pressure and syngas composition on the turbulent flame speed. Correlations were developed for ST / SL as a function of normalized parameters representing the effects of turbulence intensity, integral length scale, pressure, and temperature.

The increase of ST/SL with increasing pressure and H2 content was attributed to the thermo­diffusive and hydrodynamic instabilities. Venkateswaran et al. (2011) reported measurements of global turbulent flame speeds using a Bunsen burner, and examined the effects of Ф, syngas composition, mean flow velocity, and turbulence intensity. Consistent with other studies, the flame speed was found to exhibit sensitivity to fuel composition over a wide range of turbulence intensity, increasing significantly with the increase in H2 content. The data were further analyzed to develop flame speed correlations, indicating the effects of thermo-diffusive instabilities through negative Markstein lengths.

A normal experiment started with preheating the grate and the combustion chamber using a propane torch placed under the grate. When the temperature in the combustion chamber (2 cm above the grate) reached 800°C (after ~2 hours), the torch was turned off and biomass was added to the gasifier. The addition continued until the bed height attained 17cm; afterwards, the fuel port was closed and the flows of steam and air were adjusted to the desired experimental conditions. As the biomass was pyrolyzed and the char was burned the bed height started decreasing and the ash accumulated. Thus, biomass was added every 10 minutes and in batches as required. In the earlier batch experiments reported by Priyadarsan et al. (2004), there was no ash disposal system; as such the temperature peak moved towards the bed surface due to ash accumulation at the bottom. In the current experiments, the ash was disposed of continuously and a quasi-steady state was assured by maintaining the peak temperature at the same location in the ash disposal system. When the peak temperature achieved a steady state (~1.0 hours) the gas sampling unit was turned on and the gas analysis was performed continuously during 20 minutes by the mass spectrometer (MS).

The flowrate of dairy biomass was maintained constant at 1 kg/h and the flows of air (0.56-2.26 SATP m3/h (standard ambient temperature and pressure meter cube per hour)) at 15°C and steam (0.19-0.43 kg/h) at 100°C were changed in order to obtain the desired experimental conditions: ER = 1.59, 2.12, 3.18, 4.24, and 6.36 and S:F = 0.35, 0.56, 0.68, and 0.80. An air drier was used to dry the air before it was supplied to the gasifier. The gasifier was operated at 98 Pa vacuum pressure during all the experimentation. Temperatures along the gasifier were monitored at every 60 seconds by type K thermocouples located at 0.02, 0.04, 0.07, 0.13, 0.20, 0.24, and 0.28 m above of the grate. Samples were taken at the top of the gasifier at the rate of 0.14 SATP m3/h and conditioned by the sampling unit in order to remove tar and particulate material. The mole fractions of CO2, CO, CH4, C2H6, O2, H2, and N2 were measured every ten seconds by the MS. The same procedure used for the air gasification was again employed for enriched air gasification with little changes.

To evaluate which conversion process is more suitable for different biomasses, a preliminary characterization is necessary to analyze their chemical, physical and energetic properties.

Biomass is composed of water, ashes and dry matter without ashes and only the latter component is interesting for energy conversion yielding a calorific value. Ashes and water decrease the commercial value of biomass because:

• they decrease the bulk energy content of biomass;

• moisture absorbs energy for evaporation;

• ashes have to be disposed of;

• light ashes are transported by flue gases and contribute to PM (particulate matter) emissions;

• low melting point ashes foul heat exchangers.

Given the presence of these three different main components, the measurable quantities contained in a biomass can be expressed (Fig. 5.2): [6]

Ultimate analysis is defined as “the determination of the elemental composition of the organic portion of carbonaceous materials, as well as the total ash and moisture” (Miller and Tillman, 2008; ASTM D 5373-02; Milne et al., 1990).

The basis for the fluidized bed reactor configurations is the principle of fluidization. By forcing a gas stream (fluidization medium) through a reactor, the fuel together with the inert bed material will behave like a fluid, if the flow velocity is high enough. Air, steam or steam/oxygen mixtures are examples of commonly used fluidization media. Silica sand is the most extensively used bed material, but other bulk solids, especially those exhibiting a catalytic activity, such as olivine sand and dolomite, are also employed.

Fluidized beds provide many features not available in the fixed-bed types, including high rates of heat and mass transfer and good mixing of the solid-phase, which means that reaction rates are high and the temperature is more or less constant in the bed.

Depending on the velocity of the fluidization medium, the fluidized bed gasifiers may be divided into two categories, bubbling fluidized bed (BFB) gasifiers and circulating fluidized bed (CFB) gasifiers.

3.2.2.1 The higher heating value per unit mass of fuel

The gross or higher heating values HHV for coals can also be empirically obtained by using the Dulong equation (Annamalai and Puri, 2007), namely:

HHV[kJ/kg] = 33800 YC + 144153 YH — 18019 YO + 9412 YS (3.2)

where YC, YH, YN, YO and YS are mass fractions of C, H, N, O and S.

Another relation due to Mott and Spooner is (Mason and Gandhi, 1980):

if O < 15%

HHV [kJ/dry kg] = 103.5C%+ 1418.3 x H% + 94.2S% — 145.1 x O (organic)% (3.3) if O > 15%

HHV [kJ/dry kg] = 103.5 x C% + 1418.3 x H% + 94.2 x S%

— {153.2 — 72 x O%/(100 — A%)} x O% (3.4)

Here A = ash content.

Channiwala (1992) considered over 200 species of biomass and fitted the following equation to the data:

HHV [kJ/dry kg] = 34910 YC + 117830 YC — 10340 YO — 21110 YA + 10050 YS — 1510 YN

(3.5)

The experimental data have an error of about 1.5%.

Boie empirical equation for HHV of any fuel CcHhNnOoSs (Annamalai and Puri, 2007):

HHV[kJ/kmole] = 422272 x C + 117387 x H — 155371 x O + 100480 xN + 335508 x S (3.6)

where C, H, O, N and S are the number of carbon, hydrogen, oxygen, nitrogen and sulfur atoms respectively in the fuel. The same equation can be used to determine the stoichiometric oxygen in kg per empirical kg of fuel:

Vo2 = 32 {C + H/4 — (1/2)O + S} = 32C{1 + (H/C)/4 — (1/2)(O/C) + (S/C)} (3.7)

HHV [kJ/kg] = C{422272 + 117387 x (H/C) — 155371 x (O/C)

+ 100480(N/C) + 335508 x (S/C)} (3.8)

Based on the Boie equation, the enthalpy of formation can be derived as:

h0FJ = 28752 x{C — 0.888 x H — 6.168 x O + 6.199N + 1.337 S} [kJ/kmole] (3.9)

Thus an approximate method based on the Boie heat value exists to compute hf of any empirical fuel. If only mass fractions of C, H, N, O and S are known as YC, YH, YN, YO and YS, then the higher heating value of the fuel becomes:

HHVF [kJ/kgfuel] = 35160 YC + 116225 YH — 11090 YO + 6280 YN + 10465 YS (3.10)

One can deduce the lower or net heat value (LHV) when hydrogen in water is excluded giving: LHVf [kJ/kgfuel] = 35160 YC + 94438 YH — 11090 YO + 6280 YN + 10465 YS (3.11)

Correlation for adiabatic flame temperature with ash and moisture content is shown and plotted in Figure 3.8.

Figure 3.9 shows the higher heat or gross heat value of C-H-O fuel in kJ per kg of fuel.

3.2.2.2

The higher heat value per unit stoichiometric oxygen The heat value per unit stoichiometric oxygen (vO2) defined as:

It is well known that the HHVO2 is almost constant for most fuels. For Boie equation, the HHVO2 is given as:

HHVO2 [kJ/kg of O2] = {422272 + 117387 x (H/C) — 155371 x (O/C) + 100480(N/C)

+ 335508 x (S/C)}/(32{1 + (H/C)/4 — (1/2)(O/C) + (S/C)})

(3.13)

Ignoring S and N, trace elements in fuel, Figure 3.10 plots HHVo2, in kJ per kg of oxygen as HHV/vO2 constant. It is apparent that the HHV per unit mass of O2 burned is approximately the same of about 14250 kJ/kg of oxygen (18.6kJ/SATP liter of oxygen, where SATP means at standard atmospheric temperature and pressure ) or 3280kJ/kg stoich air (3.9kJ/SATP liter of air) for most fuels. For methane, the literature states that HV per unit O2 is 13550 kJ per kg of O2 (17.7kJ/SATP liter of O2) while Boie based equation yields 13934 kJ/kg of O2. For n-octane,

the value is 13640 kJ per kg of O2 or 17.82 kJ/liter of O2 (SATP) for CH4 while Boie yields 13730 kJ/kg O2 for Octane.

Figure 3.11 plots the respiratory quotient (RQ), a term used in biological literature (Annamalai and Silva, 2011) and defined as CO2 per kmole of stoichiometric oxygen, an indication of global

warming potential) for various biomass fuels. Typically RQ is about 1 (which is same as that of glucose, C6Hi2O6) for biomass fuels.

3.2.2.3 Heat value of volatile matter

In Figure 3.11 we see how H/C relates in fat, protein, biomass and coal. If the heat of pyrolysis is neglected, the heat of combustion of the coal can be represented as a combination of the contri­bution from the volatile matter (HVVVM) and the contribution from the fixed carbon (HVCFC) in relation to their mass percentages:

HV Coal = HVv VM + HVc FC

If FC = 1 — VM as in the case of dry ash-free (DAF) coals, one can correlate the heating values of volatiles HVv to VM (Annamalai and Puri, 2007). The Volatile Matter Higher heating value (HHVVM) was calculated using:

where HHV is the as received heating value, FC% is the amount of fixed carbon present in the fuel, HHVFC is the higher heating value of the fixed carbon (enthalpy of formation/molecular weight), and VM% is the amount of volatile matter present in the fuel.

Co-combustion in large-scale power plants can lead to an overall saving of fuels in comparison to independent fossil — and biomass-fired plants. Also, it can increase the fuel flexibility and reduce investment cost. Comparing with coal, biomass is a renewable energy source, which is considered as a CO2-neutral fuel with lower emissions of SO2, NOx, heavy metals. NOx emissions could be reduced by biomass, which has a low nitrogen and high volatile content. (EUBIA, 2007; Zhang et al., 2010; Fu et al., 2009; Daniele et al., 2007). The co-combustion of coal and biomass has many advantages which can be described as follows (EUBIA, 2007; VGB, 2008):

1. Reducing greenhouse gases emission — biomass is considered as a ‘carbon neutral’ fuel in that the CO2 emitted during biomass combustion is equal to that absorbed during the biomass growing. When biomass displaces a fossil fuel, a net reduction in greenhouse gas emissions is achieved.

2. Reducing local air pollutant emissions — burning biomass instead of fossil fuel results in lower emissions of SO2 and NOx.

3. Increasing electrical efficiency — the electrical efficiency of co-combustion power plant is higher than the traditional biomass plant, which has a small scale.

4. Ensuring security of supply — there exists a wide range of usable biomass fuels. Varying qualities and quantities of fuels can be partially compensated by adjusting the co-combustion rate.

5. Reducing cost — co-combustion presents the opportunity to use the existing fossil-fuel fired power plant infrastructure, which can be modified for co-combustion relatively easily. An optimum thermal biomass blending ratio of biomass co-combustion is 10% (on an energy basis) (Munir et al., 2010). Addition of biomass to a coal-fired boiler does not impact or at worst only slightly decreases the overall generation efficiency of a coal-fired power plant. Compared with other renewable options, biomass co-firing represents the most cost-effective means of renewable power generation in many cases (Belosevic et al., 2010; Baxter et al., 2005; Hein etal., 1998).

Meanwhile, co-combustion of coal and biomass has some disadvantages shown as follows (VGB, 2008):

1. Preprocess — A fuel handling system is designed for a particular water content, size distribution, dust etc. With co-combustion of biomass it is necessary to adapt the existing or even build a new combustion system for that fuel.

2. Corrosion — Higher corrosion risk due to increased HCl formation in case of substitution of fuels with higher chlorine content (sewage sludge, some cereals). Many biomass fuels contain large amounts of alkalines, especially potassium, which may aggravate the fouling problems (Baxter et al., 1993; Bakker et al., 1997; Robinsin et al., 2001a, b; Dunaway et al., 2003; Lokare etal., 2003).

3. A SCR DeNOx catalyst can be blocked by ash particles or deactivated by potassium, chlorine, and in case of sewage sludge also poisoned by some heavy metals and metalloids (As, Zn).

4. Operating costs are typically higher for biomass than for coal. The most sensitive factor is the fuel cost. Even if the fuel is nominally free at the point of its generation (as many residues are), its transportation, preparation and on-site handling typically increase its effective cost per unit energy such that it rivals and sometimes exceeds that of coal.

For the utilization of the ash in the cement and concrete industry, the concentrations of alkali metals, P2O5, SO3, Cl and unburned carbon in the ash are the critical parameters. It was found that the ashing temperature should be selected according to the biomasses proportion, when the biomass fraction is raised, the ash fusing temperature of blends decreases generally, and biomass with high P and K content proportion should not exceed 10% in co-firing (Dong et al., 2010).

This process originates the visible flame following the reactions of flaming combustion. Oxidation happens generating a diffusive flame due to slow combustion because volatiles substances exiting form the particle and oxidizing agent are not pre-mixed.

As reported by (Miller and Tillman, 2008) during volatile oxidation different free-radical reactions happen, such as: [7]

A simple chemical equation for levoglucosan combustion is the following (Sullivan and Ball, 2012):

C6H10O5 + 6O2 ^ 6CO2 + 5H2O (5.18)

Besides the stoichiometric equation it has to be considered that several intermediate compounds are created during flaming combustion of the volatiles. Woodley (1971) has identified about 40 compounds, produced by the thermal degradation of levoglucosan.

The oxidation reactions are exothermic and very fast. The activation energy for the oxidation of levoglucosan is about 190kJ/mol, the frequency factor is about 2.55 + 1013 s—1. The reaction enthalpy for complete combustion is —14kJ/g (Parker and LeVan, 1989).

Some other gas phase homogeneous reactions are reported in the following Table 5.5.