seljak 2019 emission reduction through highly oxygenated

Cite this paper as: Seljak T, Katrašnik T: Emission reduction through highly oxygenated viscous biofuels: Use

of glycerol in a micro gas turbine, Energy 2019, doi: 10.1016/j.energy.2018.12.095.

___________________________ *

Corresponding author: [email protected]

1

Emission reduction through highly oxygenated viscous biofuels: use of glycerol in a micro

gas turbine

T. Seljak1*, T. Katrašnik1

1 University of Ljubljana, Faculty of mechanical engineering,

Aškerčeva 6, SI-1000 Ljubljana, Slovenia

*corresponding author

[email protected]

Abstract

The presented paper focuses on the combustion analysis of a highly oxygenated, viscous and economically viable fuel

in an internal combustion micro gas turbine. Results suggest that environmental benefits in terms of NOx emissions

and PM emissions are significant as concentrations of both species are reduced. Based on the analysis presented in

the study, altered local air-fuel ratio suppresses NOx formation and PM formation resulting in 10-fold lower PM

emission than same experimental apparatus with diesel fuel. The study further analyses and identifies three key

mechanisms that are responsible for this reduction: altered oxygen profile during mixture preparation, prevention of

soot precursors formation and promotion of soot oxidation reactions. The influence of tested fuel is also perceivable

through higher CO and THC emissions. Parametric analysis within the selected experimental space suggests that

operational parameters of the micro gas turbine can be adjusted to reduce concentrations of these two species without

negatively affecting other pollutants. Significant benefits of the fuel in terms of environmental impact indicate that

further investigation and development of this approach might provide a new alternative energy source for stationary

use as well as for mobility in sectors where heavy duty gas turbines are used as prime movers.

Keywords: viscous biofuel, glycerol, oxygenated fuels, particulates, soot, emissions

1. Introduction

Recently, significant legislative push led to wide popularization and investments in the technology of 1st

generation biofuels. As a result, demand for oils and sugars increased to the point where these fuels started to compete

with food feedstock, raising the prices of arable crops. To offset this negative effect, further stimulation of 1st

generation biofuels is limited and efforts are made to displace them by subsequent generations of biofuels [1].

However, significant production capacities, particularly of biodiesel are still present and are expected to stay around

since up to 7% of FAME is allowed in the diesel fuel as specified in EN590 and RED 2009/28/EC dictates 10%

renewables content in transport fuels by year 2020 [2]. Under these terms, the largest producer of biodiesel is EU,

responsible for 65 % of world’s biodiesel capacities. Consequently the quantities of by-product - residual glycerol (or

crude glycerol) (GLY) are higher than ever [3] and the prices range from 150 – 250 €/t. Comparing it to most common

energy sources, crude glycerol is is positioned among one of the cheapest energy carriers available:

0,04 €/kWh – natural gas

0,08 €/kWh – electricity

0,03 €/kWh – IFO 180 (bunker fuel)

0,04 €/kWh – crude glycerol

Besides using it as a basic chemical (for food products, personal care products, polyethers/polyols,

pharmaceuticals,…), what usually requires significant purifying or upgrading procedures, an attractive option is also

to use it as an energy carrier as the requirements for purity and contaminant content could be less strict in this case,

particularly if robust and heavy duty applications are considered. Ideally, to boost the energy conversion efficiency

and allow for the use of glycerol on a wide variety of power output scales, it is advantageous to rely on the internal

combustion concepts as they provide the necessary flexibility and maturity of technology.

Combination of internal combustion and physical and chemical properties of the glycerol, namely its high

oxygen content and high viscosity [4] imposes significant technical challenges in terms of application in combustion

systems with high power density which is linked to relatively small combustion volumes and residence times. As

modern piston engines are designed with more stringent requirements on physical and chemical properties of the fuels,

gas turbines can be considered as preferred internal combustion engine type for glycerol combustion. Gas turbines

namely feature continuous combustion, long residence times of the combustible mixture and possibility to use air-

assisted atomization nozzles, which are advantageous when using highly viscous and low-volatile fuels [5–7].

So far few studies were dealing with combustion of glycerol and majority of them included atmospheric

continuous combustion devices [8–10]. These studies provide valuable experience in the area of combustion of highly

https://doi.org/10.1016/j.energy.2018.12.095



2

oxygenated fuels. Possible upgrade to this approach is implementation of atmospheric combustion system to Rankine

cycle or to externally fired joule cycle [11]. The present study takes this to even higher level of complexity by

investigating the direct combustion process of glycerol in a micro gas turbine (MGT) and by this representing, to the

best of author’s knowledge, the first study in this area. The advantage of this approach is that direct conversion to

mechanical energy is possible in the efficiency range of 30-37% [12] and the outcomes of the study are transferrable

to units for decentralized stationary power generation or to marine applications.

Work done on gas turbines with similar fuels is mostly limited to biomass pyrolysis oil/ethanol blends [13]

and studies with MGT combustion of pure liquefied wood, a fuel similar in properties to pyrolysis oils, where several

studies were published with stable combustion albeit significant technical challenges linked to its high viscosity

[5,6,14,15]. Both fuels feature nearly 50% oxygen content. These studies can be used as a basis for a present work

since they partly describe the approach suitable for the use of highly viscous oxygenated fuels in MGTs. Systematic

guidelines for such highly viscous fuels are also separately described in papers dealing with appropriate workflow

[16,17] and also specific component and control features that are required in such cases [16,18].

Benefits that oxygenated fuels have on particulate matter (PM), carbon monoxide (CO) and nitrous oxides

(NOx) emissions were identified in several studies on piston engines, which in majority investigated lightly oxygenated

multicomponent fuels or their dual-fuel use. Examples of such blends are Nitromethane-Diesel-Butanol blends [19],

gasoline-methanol in dual fuel engines [20], biodiesel/butanol, biodiesel/DMF and biodiesel/ethanol in RCCI engines

[21] as well as diesel/biodiesel/ethanol/DMC in numeric models of diesel combustion [22]. Such studies indicate that

combustion process can be highly influenced by the presence of oxygenated species regardless of combustion device

and process parameters, however the phenomena through which these species are affecting the combustion kinetics

can significantly differ among the investigated cases. The phenomena that potentially occur during MGT combustion

of highly oxygenated fuels will be given hereinafter

1.1 Objectives

The main objective of this study is therefore to set up a first successful combustion of technical glycerol in a

MGT and to identify the mechanisms that are influenced by high oxygen content of the fuel and profoundly influence

the emissions of gaseous species and particulate matter. It is further aimed that the results will provide a solid basis to

exploit the full potential of highly oxygenated biofuels while at the same time addressing also topical matter of low-

grade secondary materials and waste streams.

This will be achieved through several specific objectives that separately cover the key areas of combustion

and emission formation phenomena:

- First, analyses of thermodynamic responses of a MGT using glycerol as a main fuel and diesel fuel for

benchmark data will be performed to confirm the comparability of results.

- Parametric analysis of sensitivity of CO and THC emissions on operational parameters of the MGT will be

performed to asses the possibility of concentration reduction of these two species.

- Further, specific analyses of NOx emissions and PM emissions, which are most profoundly influenced by

oxygen content of the fuel, will be elaborated.

- Key underlying phenomena for reduction of NOx and PM emissions when using oxygenated fuels will be

identified and discussed to provide a basis for further improvement of the combustion process.

-The influence of oxygen presence in the fuel will be elaborated in terms of mixture formation process and

oxygen concentration profile around the fuel nozzle.

- Brief discussion on possible contaminants and their effects will be given along the economic viability of

crude glycerol use for the purpose of power generation.

- Economic feasibility of crude glycerol use will be evaluated based on sensitivity calculations of Return of

Investment interval.

2. Methodology

To overcome the significant technical difficulty linked to glycerol viscosity its reduction is mandatory to obtain

sufficient quality of atomization, which is capable of supporting combustion process under conditions present in MGT.

Considering the performance of the used experimental systems it is foreseen that innovative solutions linked to the

injection components will allow the use of crude glycerol without blending or upgrading and only by preheating the

fuel. To do this, the full list of recommended adaptations that proved useful for utilization of highly viscous fuels was

adopted [16] in a laboratory scale MGT. Particular importance is laid upon thermal protection of the nozzle, which

prevents development of carbon deposits that are a limiting factor in increasing the fuel temperatures [16]. Key data

of the tested GLY is presented in Table 1, together with properties of benchmark diesel fuel according to EN590:2011

(D2). The molecular structure of GLY is presented in Figure 1.




3

Figure 1: Glycerol molecule C3H8O3.

Table 1: Properties of analysed fuels.

GLY D2

C 42.19 87.00

H 9.14 13.00

N 0 /

S 0 <0.001[23]

O (by diff.) 48.67 /

Density 1.26 kg/L 0.820 –

0.845 kg/L

[23]

LHV 19.0 MJ/kg 42.2 MJ/kg

Stoichiometric

ratio 5.19 14.7

Viscosity at 70 °C 50.6 mPa.s 1.3 mPa.s




In terms of viscosity, it is currently considered that 15 mm2/s [24] is the maximum value for gas turbines, although

some authors propose even lower values 12 mm2/s [25] or 10 mm2/s [26]. However, these values are linked to pure

pressure atomizers, which are known to have limited turndown ratio and are highly sensitive on viscosity of the fuel

as the energy for liquid jet break up should be delivered solely by fluid pressure. More suitable nozzles for high

viscosity fuels are therefore twin-fluid atomizers which feature better turn down ratios (i.e. in certain interval of mass

flows, the atomization ability is minimally influenced) and are less sensitive to viscosity as energy for liquid jet

breakup is delivered also by atomizing air stream [27]. However, the viscosity levels of GLY are too high to be

accommodated even by twin-fluid nozzles.

According to temperature dependence of GLY viscosity (Figure 2), the acceptable levels can be obtained by

preheating the GLY to pre-specified temperature. Useful temperature interval was for the present study constrained

by data, obtained in the literature [5], where fuels with high viscosity were fired in a similar experimental system.

Selected interval covers temperatures between 50 – 70°C, yielding viscosities according to Table 1.

Figure 2: Temperature dependence of GLY viscosity.

10

100

1000

10000

10 30 50 70 90

Vis

cosi

ty [

mP

a.s]

T [°C]




4

The test rig constructed here consists of a directly fired experimental gas turbine, a fuel conditioning system and

emission analysis equipment. The combustion chamber was of single can design. Single stage compression was

provided by a radial compressor powered by a radial single stage turbine. Details on the gas turbine setup are also

presented in [5] and Figure 3. A wide range of operating parameters were achievable by operating the turbine across

several different turbine inlet temperatures (TIT) which resulted in:

• different equivalence ratios (EQR) due to different enthalpy extraction on a turbine and different pressure

drops on flow path,

• slightly different primary air temperatures as a result of different turbine outlet temperatures and different air

mass flows.

Since the experimental setup is designed purely for combustion analyses, power output is neither optimized nor

directly measured. Instead, the power withdrawal is emulated by the throttle valve in the turbine discharge duct. The

pressure drop over the throttle valve increases with increased fuel flow and by this proportionally decreases the

enthalpy difference on turbine rotor. With this approach, similar thermodynamic conditions can be obtained in

combustion chamber as with small variations of power output on the turbine shaft in regular MGTs [27]. Although

low, power output can be estimated by mentioned enthalpy differences (measured by pressure difference, temperature

difference and mass flow over throttle valve in different operating points. Power derived from this data for GLY is

given in Table 2 and is calculated based on the equation (1).

𝑃𝑐𝑎𝑙𝑐 = (�̇�𝑓𝑢𝑒𝑙+�̇�𝑎𝑖𝑟)(𝐻5(𝑇, 𝑝) − 𝐻4(𝑇, 𝑝))

Where variables represent the following:

�̇�𝑓𝑢𝑒𝑙 fuel mass flow

�̇�𝑎𝑖𝑟 air mass flow

𝐻5(𝑇, 𝑝) enthalpy of gas downstream of throttle valve

𝐻4(𝑇, 𝑝) enthalpy of gas upstream of throttle valve

The design of injection nozzle followed the commercially available solution, however the nozzle discharge

surfaces were modified in order to accommodate the thermal protective layer. By this the radiative and convective

heat transfer was reduced to a point where no fuel decomposition occurred in the internal mixing chamber of the

nozzle and formation of deposits on discharge surfaces was minimized. This is a key improvement which enables high

preheating temperatures which otherwise cause excessive fuel coking issues as was already reported in [6] and [27].




5

AVL fuel balance0 – 150 kg/h

Uncertainty 0.12 %

D2 reservoir

Micromotion CMFS 1 – 310 kg/h

Uncertainty 0.05 %

GLY reservoir

Heating vessel

T

P

T

P

T

P

T

P

T

P

TP

T

P

Compressor Turbine

Combustion chamber

Recuperator

Exhaust diversion valve

Throttle valve

Pressure sensor; 0 – 10 bar; Uncertainty 0.05 %

Temperature sensor; K type; Uncertainty 1.0 %

Meriam airflow meter

Uncertainty 0.86 %

Emission analyzersNDIR cell

CLD detectorFID detector

Heated line 195°C

Exhaust in regenerative cycle mode

Exhaust in simple cycle

mode

Air intake

Figure 3: Experimental system layout.

The emission equipment consisted of a wet exhaust gas analyser performing separate consecutive measurements

of total hydrocarbons (THC) in an FID cell, nitrous oxides (NOx) in a CLD cell and CO, CO2 and H2O in an NDIR

cell.

Particulate matter (PM) was analysed using photoacoustic method, suitable to detect particles with high

absorption coefficient. Considering that the glycerol featuring technical quality with no expected ash precursor content

was used in the study, the emitted particles will most likely contain mainly carbon, giving the black appearance and

high absorption coefficient of the PM. In this view, no underestimation of soot emission should occur although this is

relatively common when using fuels with low purity and relying on the mentioned method of PM measurement.

The experimental matrix was designed with integrity of obtained data in mind. The influence of viscosity was

emulated through variation of fuel temperature and power output was emulated through variation of turbine inlet

temperature. Key characteristic operating points where data was obtained are presented in Table 2.




6

Table 2: Key operating points for gathering the emission data.

Target Turbine

inlet temperature

(TIT) [°C]

Series name

Primary air

temperature

[°C]

Actual

TIT [°C]

Air mass

flow [kg/s]

Fuel mass

flow [kg/h]

Calculated power [kW],

based on

(ΔH [kJ/kg])

750

GLY-70°C

437.0 752.5 0.154 17.8 3.66 (23.1)

800 461.5 804.3 0.170 21.2 4.19 (23.9)

850 493.7 854.3 0.179 23.4 4.87 (26.3)

900 523.7 910.7 0.190 27.2 5.74 (29.1)

750

GLY-60 °C

440.9 748.2 0.144 15.8 3.69 (24.8)

800 465.2 804.9 0.164 19.9 4.53 (26.8)

850 489.4 848.1 0.174 22.6 5.00 (27.7)

900 515.8 902.3 0.187 26.6 5.75 (29.6)

750

GLY-50 °C

439.8 741.9 0.141 16.4 3.44 (23.6)

800 475.0 811.1 0.163 20.1 4.54 (26.9)

850 498.6 857.8 0.176 23.3 5.08 (27.9)

900 530.4 904.2 0.185 25.7 5.62 (29.3)

750

D2-20 °C

445.0 763.0 0.141 5.6 3.41 (23.9)

800 478.5 826.3 0.146 6.2 4.47 (30.2)

850 499.3 853.9 0.164 7.3 5.08 (30.5)

900 522.0 904.6 0.182 9.1 5.81 (31.5)

3. Results and discussion

The following sections present the results of key operating parameters under different operating conditions in

dependency to TIT. This parameter is directly linked to EQR and temperature of combustion chamber intake air

(PAT). Similarly, the pressure ratio increases with TIT. The presentation of thermodynamic parameters is followed

by the presentation of the concentrations of emission species (CO, THC and NOx) over different TIT, supported by

the explanation of underlying mechanisms of formation. Finally, PM emissions are presented.

3.1 Thermodynamic parameters

Figure 4 presents the equivalence ratios (EQR) for different operating points. As low fuel mass flows were

required to obtain the desired TIT, EQRs are relatively high. This is a consequence of high degree of exhaust heat

regeneration, leading to high primary air temperatures as discernible in Figure 5. This comparison is revealing that

although GLY and D2 are inherently different fuels (different heating value, stoichiometric ratio and density), the

baseline thermodynamic parameters can be fully replicated with GLY, providing equal energy inflow by the fuel. This

means that fuel mass flow needs to be increased to compensate differences in fuel heating value. This also indicates

that there are no significant chemical losses in terms of unburnt fuel (THC and CO emissions), which will be further

confirmed through emission measurements in the following sections. At this point it has to be noted that local

equivalence ratio in primary zone of the combustion chamber varies between 0.6 and 0.8 as approximately 70-85% of

the overall airflow is diverted past primary zone to secondary and dilution zone. The majority of fuel oxidation and

thus heat release therefore occurs at temperatures higher than TIT as is common for combustion chambers in gas

turbines. This enables sufficiently high activation energy for dissociation of components with high auto-ignition

temperature (GLY~ 370 °C, D2 ~ 210°C).




7

Figure 4: Equivalence ratio for GLY and D2.

Figure 5: Primary air temperatures for GLY and D2.

3.2 Sources and mechanisms responsible for CO and THC formation

Figure 6 presents CO emissions at different turbine inlet temperatures (TIT) for both fuels and for all tested

temperatures. Based on the physical and chemical characteristics of the fuels presented in Table 1 and Figure 2,

particularly the viscosity and density of GLY and to some extent also its evaporation curve, the higher CO emissions

of the GLY in Figure 6 are somewhat expected. Nearly 10-fold higher viscosity of GLY compared to D2 impairs the

atomisation ability of this fuel, resulting in a longer life span and penetration depth of the larger and denser droplets.

Increased penetration depth of the droplets is the consequence of decreased surface to mass ratio of the droplets due

to larger initial droplet diameter and due to the increased density of GLY. Furthermore, delayed evaporation of

droplets (which in this case feature lower surface to volume ratio due to larger diameters) and unfavourable

evaporation curve of GLY in comparison to D2, requiring higher temperatures to fully vaporize the fuel is in this case

causing prolonged time interval in which the mixture preparation takes place. As the mixture is formed relatively late

in the primary zone of combustion chamber, the time available for combustion reactions before the dilution with air

in dilution zone of combustion chamber can cause excessive reaction quenching. Furthermore, high auto-ignition

temperature (370 °C) and high boiling point (290°C) of the GLY can also play a significant role (by offsetting the

ignition of the mixture towards the end of primary recirculation zone leading to a shift of reactive zone towards the

secondary zone of combustion chamber. These phenomena inevitably lead to higher CO emissions in exhaust gas

stream compared to the D2 fuel and could generally be avoided by decreasing the duration of mixture preparation or

by elevating the temperature levels in primary zone and secondary zone of combustion chamber.

Another mechanism contributing to higher CO exhaust emissions of GLY is the entrapment of mixture, which

generally has longer penetration length, into liner cooling air where aforementioned quenching occurs even earlier.

This phenomena is strongly linked to survival of GLY droplets and could be avoided either by careful optimization

of spray pattern or by approach discussed above.

Despite the fact that fuel properties indicate shift of the main reactive zone deeper downstream in the combustion

chamber due to the delayed evaporation process, a surprisingly large volume of the flame was also observed near the




8

fuel nozzle. This most likely related to high bonded oxygen content and high degree of premixing of the GLY in the

internal mixing chamber of the injection nozzle. Bonded oxygen namely accounts for 35% of the stoichiometric

oxygen, hence lowering the need for external air entrampent in order to reach the flammability limits of the GLY.

This can, in the case of high primary air temperatures, cause relatively early ignition of small amount of mixture which

is in a suitable range of EQRs. At the same time, phenomena linked to larger droplets still persist, thus the penetration

of the part of the mixture could be sufficient to reach the colder parts of combustion chamber (liners or secondary

zone). This earlier ignition of GLY in comparison to D2 is further elaborated in the following sections, where the role

of bonded oxygen on EQR iso-surfaces is analysed.

Figure 6: CO emissions of D2 and GLY.

Figure 7. THC emissions of D2 and GLY

Considering upper phenomena resulting in higher CO emissions with GLY in comparison to D2, it was presumed

that increasing TIT would lead to suppression of listed phenomena by increasing the temperature levels in combustion

chamber and by this increasing the heat transfer to the droplets and consequently also evaporation rate of the fuel.

This hypothesis was confirmed as increasing the temperature levels resulted in reduction of CO emissions, since the

reduction is nearly 10-fold within the operation area of the MGT along different TIT as seen in Figure 7.

Another potential measure to speed up the mixture preparation process is increasing fuel temperature. This

approach is mainly affecting an earlier stage of spray formation process, namely the primary and secondary spray

break-up which, followed by reduction in viscosity of glycerol due to higher temperature, yield smaller SMD of

primary as well as secondary droplets. Higher surface/volume ratio of droplet cloud enables higher evaporation rate,

whereas higher fuel temperature also reduces the required temperature margin for heating up the fuel to the point of

GLY boiling temperature (290 °C). The reduction of CO emissions in terms of highest tested fuel temperature (70°C)

is mostly observable at low TIT, where the contribution of faster evaporation and mixture formation is the most

pronounced, yielding roughly 3-fold reduction compared to 50°C fuel temperature.

Similarly as CO emissions, THC emissions in Figure 7 are also strongly TIT-dependent and fuel temperature

dependent for GLY, which is in line with phenomena discussed for CO. Also here, the difference between D2 and

GLY is significant, while relative differences in emissions between both fuels are smaller in comparison to CO

emissions. When looking at complexity of both fuels in terms of their molecular composition, it can be noted that




9

GLY is significantly simpler fuel with its single-component composition. On the other hand, D2 generally contains

several thousand hydrocarbon components with different structures and thus its high temperature oxidation reactions

contain significantly larger number of steps and interactions before generating roughly the last intermediate – CO. In

the case of GLY this stage of combustion requires fewer steps and thus CO is most likely formed after involving less

intermediate products and faster in comparison to D2, thus THC emissions are not significantly higher than for D2

since few intermediates of GLY are oxidized relatively fast. Based on this observation, the combustion of GLY should

generally be less technically demanding, however all of these advantages are offset by the unfavourable auto-ignition

temperature and challenging mixture preparation process together with different evaporation curve which still sum up

to higher THC emissions than D2.

It is foreseen that careful adjustment of operational parameters in terms of using high TIT, regenerative cycle

MGT with high primary air temperature and high fuel preheating temperature could reduce the CO and THC emission

to the levels, comparable to D2. As will be shown in the following sections, this adjustments would not significantly

affect the elevation of NOx and PM emissions as the NOx-PM trade off is not very pronounced in the case of GLY.

3.3 Sources and mechanisms for NOx formation

Figure 8: NOx emissions of GLY and D2.

NOx emissions concentrations reported in Figure 8 demonstrate two distinctive trends. The first is linked to fuel

type and the other one to TIT which is related also to primary air temperature and EQR. The highest NOx emissions

are observed with D2 and high TIT, followed by GLY and high TIT.

As no nitrogen is present in both tested fuels, the NOx emissions are a consequence of a thermal (Zeldovich) and

prompt (Fenimore) mechanism. The NOx component formed through N2O can be neglected in presented cases as

pressure ratios are low due to single stage compression (1,9 – 2,5) and reaction pathway for N2O requires three-body

reaction, associated with high pressures. The general trend of increasing NOx emissions of GLY and also D2 with TIT

can clearly be attributed to increasing rate of thermal NOx formation. There is little to no influence of GLY preheating

temperature so no pronounced trend can be observed in this case. The same dependence on TIT, although slightly

more pronounced is also observed with D2. The effect of temperature levels on NOx formation is therefore similar for

both fuels. Further elevation of TIT and GLY temperature would manifest itself in elevation of NOx, but in a range of

few ppm so this would indeed be a successful measure to further reduce CO and THC in the section above.

Nevertheless, a 5-fold difference between absolute NOx concentrations of D2 and GLY is a bit unexpected at a first

glance. Low levels of NOx emissions in combustion of GLY might be the consequence of high oxygen content and

low stoichiometric ratio in GLY. Therefore the amount of required air to reach a mixture of GLY and air inside a

flammability interval is smaller. Consequently also lower amount of nitrogen is delivered into high temperature zone

thus reducing the possibility for thermal NOx formation. Temperatures on the rich side of the flame are due to sub-

stoichiometric conditions relatively low as are the nitrogen concentrations, which leads to low formation rate of NO.

The transition of the partially reacted mixture, which is at this stage already diluted with products of partial oxidation

and relatively small amount of nitrogen, to lean conditions reduces the temperatures and prevents further formation

of NO or even its dissociation. This phenomenon resembles the approach of partially premixed combustion (for

example RQL, where rich mixture is purposely established in the initial phase of combustion and is after ignition

swiftly diluted to lean mixture thus avoiding the occurrence of high temperature combustion zone).

At this point it is important to clarify that combustor design in present study is purely diffusive with single point

injection and partial premixing is occurring mostly on micro and molecular level without dedicated flow pattern to

induce these conditions. The premixing is achieved mainly with bonded oxygen, however the addition of atomizing




10

air from the injection nozzle might play a minor role, but this effect is also present with D2, so relative effect to GLY

can firmly be attributed to the described phenomenon. The oxygen content in the fuel, although it has several other

disadvantages, has a significant potential to reduce the NOx emissions by reducing the stoichiometric air ratio. This is

clearly valid only for fuels where low stoichiometric ratio is a consequence of high oxygen content (therefore to fuel

O/C ratio) and not to fuels where low stoichiometric ratio is caused by inclusion of inert components (ash, nitrogen,

water…). In the latter case, the reduction of NOx is not very well pronounced and is mainly a consequence of altered

heat capacity of reactants or combustion products. Discussed oxygen concentration conditions with GLY are also

presented in Figure 11 and the identified mechanism of NOx reduction is also provided in the summary in (Path D in

Figure 10) as it is closely linked also to PM emission reduction.

3.4 Sources and mechanisms for PM formation

The applied method for measurement of PM relies on absorption coefficient of PM in the sample stream which

is fitted to high absorbance of carbon material (soot). Detection of particles with significantly lower absorption

coefficient than soot is therefore limited. The results in Figure 9 therefore roughly correspond to soot emissions

although minor false positives of soot emission might occur in case other particles are present – the quantity of these

is estimated to be low in the case of GLY, as no ash was present in it.

Figure 9: Soot emissions of D2 (right axis) and GLY (left axis)

The soot emissions from D2 are roughly 10-fold higher than for GLY and exhibit a pronounced dependence on

TIT for D2. This behaviour is mostly linked to increasing EQR with increasing TIT (as seen in Figure 4) which shifts

the primary zone EQR towards 0.8, where local areas within soot formation interval are more likely to occur (related

to temperature and EQR). One of the most interesting observable trends is the soot emission of GLY which are

surprisingly low considering its physical and chemical properties. High viscosity, density and low heating value

accompanied by high boiling and high auto-ignition temperature of the fuel are usually linked to high sooting

tendency. In practice, the main mechanisms for this is the diffusive pathway for soot formation due to larger droplets

and aggravated oxygen diffusion to fuel rich areas which then often exhibit the required conditions for soot formation.

It seems that this is not the case with GLY as the negative effect of higher viscosity is not really pronounced. This is

visible through low dependence of soot emissions on fuel temperature, which is still yielding notable changes in

viscosity. These changes cause only minor reduction of soot emissions which are believed to be a result of lower

number of large droplets which occur more frequently at lower preheating temperature of GLY (and higher viscosity).

These are capable of reaching the quenching zone of the combustion chamber where the pyrolysis process occurs and

yields small coke fragments, perceived as soot.

The soot formation threshold generally depends on temperature, EQR and pressure [28]. It is widely accepted

that temperatures in the range 1350 K - 1900 K and EQR in the range 2,9 – 1,8 [29] are prerequisite for soot formation.

As the pressure differences between different operating points of the experimental turbine are relatively low,

temperature and EQR are the main driver for the observed differences, which are both interrelated and pressure

influence can be neglected at this stage of research.

The TIT dependence of soot from GLY is exhibiting the opposite trend as D2 where soot emission reduces with

TIT. As the energy density of the combustible mixture is comparable to D2 and the adiabatic flame temperature of

GLY is close to that of D2, the influence of TIT on soot formation with GLY should be similar to that of D2. At this




11

point, three possible underlying mechanisms can be identified that are involved in prevention of soot formation or its

faster oxidation kinetics. The identification of these mechanisms is important in terms of understanding and further

optimization of combustion process of oxygenated fuels. Mechanisms are discussed below for the current case of GLY

combustion. To assure the transferability of results of this study the three key identified mechanisms are then provided

in the next section without including GLY specific phenomena.

The first reason (path A in Figure 10) for the surprising PM emission trends can be linked to high oxygen content

in GLY and purely physical phenomena involved in spray formation. Bonded oxygen becomes during dissociation of

GLY readily available in the flame zone in the form of free radicals, resulting in high local C/O ratio already without

any entrapment of external air. Thus in case of GLY the C/O ratio is always below 1 (based on elemental composition

of GLY), regardless of the mixture preparation dynamics, whereas for D2 this value is easily exceeded on the rich

side of the mixture, where insufficient air is entrapped into the spray and approaches infinity where only fuel vapour

is present. Areas outside the soot formation interval are more likely to occur with GLY as volume of the mixture

within the soot formation limit (EQR between 1,8 and 2,9) is smaller. Additionally, the high O/C ratio of the fuel also

leads to a much earlier onset of exothermic oxidation reactions, even when a very low quantity of air is available in

the fuel-air mixture. The temperatures in the fuel rich areas could also be elevated above the soot formation threshold

when using GLY [30]. In contrast to GLY, in the case of D2, oxygen is not readily available for the early onset of

exothermic reactions and the temperatures in fuel rich areas are lower, thus enabling soot formation. The temperature

and EQR profile versus the O/C ratio are therefore the main reasons for the significantly reduced soot emissions for

GLY. This is further elaborated also in Figure 11, where lower and upper soot formation threshold is provided for

GLY and D2.

Second possible mechanism (path B in Figure 10) that might be accelerated with high oxygen content is a

suppression of a widely accepted acetylene hypothesis of soot formation which might not play a significant role in

case of GLY due to OH radicals disrupting the formation of acetylene by limiting aromatic ring growth [31]. It has to

be noted that each C atom in GLY has its OH counterpart due to three oxygen molecules being available within the

GLY. Formation of PM in the case with GLY most likely relies on less known diffusive hypothesis for soot formation

mechanism, which relies on pyrolysis as the main process that causes soot formation [32]. Furthermore, the availability

of O and OH radicals in the flame zone promotes carbon oxidation directly to CO and CO2 without significant

recombination of intermediate products, especially when noting that C-O bond strength is greater than that of C-C

bond, thus glycerol combustion pathway might strongly prefer the generation of C-O species over C-C species which

act as soot precursors. Thus, glycerol specific behaviour can contribute to generation of shorter soot precursors (up to

C3) which again prevents the formation of complex hydrocarbons [31].

The third mechanism (path C in Figure 10) is linked to post formation behaviour of soot. According to studies

performed in this area [33], the soot, originating from oxygenated fuels often could contain different oxygen groups

on the surface which might originate from partially reacted or recombined fuel molecules. These oxygenated species

are highly reactive and serve as initial reactive points for soot oxidation. Speeding up of oxidation process happens

by generation of a porous soot structure which spreads from these initial reactive points and then lead to consequent

internal oxidation of soot. The hollowing out of the soot particle thus generates significantly lower soot mass and

particles with large active surface, which is essential soot burn-off with surface reaction of fixed carbon to CO and

CO2.

Similar observations with very low soot were already made in [10], where GLY was combusted in an atmospheric

furnace and in [6,34] where high preheating temperatures of similar highly oxygenated fuel also reduced soot

concentration below measurable limit. The present study is confirming that this benefit can be transferred also to

internal combustion engines, which might open the way for further reduction of PM and NOx emissions in

decentralized manner or in marine mobility applications where engines are already technically sufficiently advanced

that can utilize highly viscous fuels. Steps in this direction have already been made by trials of heavy fuel oil / GLY

blends which provided promising results, however the present study indicates that for obtaining a full benefit in terms

of conversion efficiency, costs and environmental benefit, pure glycerol can be used in applications where gas turbines

are used as prime movers. At the same time this would help to avoid the need for producing emulsions as in case of

heavy fuel oil / GLY blends. This would greatly reduce the technical disadvantages linked to the unstable blends of

polar and non-polar fuels.

3.4 Summary of mechanisms responsible for NOx and PM reduction

To ensure the transferability of results also tu other oxygenated fuels, Figure 10 summarizes the mechanisms

responsible for low emissions of NOx and PM when using highly oxygenated fuels. The list of mechanisms is

constructed from analysis of combustion process in this study and by the support of studies available in the area of

oxygenated fuels [6,8,9,33,35,36] and provides a basis for further improvement of combustion process. The

mechanisms are generalized in comparison to the discussion above so they are transferrable also to other oxygenated

fuels and they do not contain GLY specific reasoning.




12

Figure 10. Mechanisms responsible for PM and NOx reduction during combustion of oxygenated fuels (specific

phenomena are provided in grey box).

As described in earlier sections an important correlation is observable between PM and NOx (path A to path D in

Figure 10) emissions as the first mechanism (altered oxygen concentration profile during mixture preparation) also

significantly contributes to the reduction of NOx emissions. As this mechanism is probably the most important for

both components, PM and NOx, a further insight is necessary. This can be done by exposing the conditions dictated

by local EQR through implementation of an innovative mechanistically based 0D spray model [37] which calculates

the local EQR in spray cloud and takes into account fuel stoichiometric ratio. The main assumptions, important for

interpretation of calculated data in the presented case are:

The momentum in-flux is only through the fuel injected from the nozzle

The spray is circumferentially symmetrical along the axis of the nozzle.

At any distance from the nozzle, the spray velocity exhibits a Gaussian distribution around the centerline

By taking into account equation (4) [37], it is possible to calculate the position of iso-surface that represents the

same EQR in the spray cloud.

𝑟𝜆,𝑥 = 𝑅

√𝑙𝑛

(

𝑥tan(

𝜗2)

√2𝑘

𝜋𝜌�̇̃��̇̃�𝑓(1+𝜆𝐿𝑠𝑡)

)

−𝑘 (2)

where

𝑅 = 𝑥tan (𝜗

2), (3)

and

�̇� = ∫ 𝜌𝑢2𝑑𝐴𝑅

0, (4)

and variables represent the following:

𝜗 spray cone angle (deg)

𝑘 constant in the Gaussian profile (-)

𝜌 density (kg/ m3)

Altered oxygen

concentration profile

during mixture

preparation.

Prevention of soot

precursor formation.

Promotion of soot

oxidation reactions.

Mechanisms influencing the PM emission formation in oxygenated fuels

Mechanisms influencing the NOx emission formation in oxygenated fuels

High local C/O ratio in

rich regions of the spray

core due to oxygen

presence in the molecular

structure of the fuel.

High concentration of O

and OH radicals within the

flame and post-flame

products.

Presence of initial oxygen

groups on the surface of

generated soot.

High local EQR ratios

reduce the volume of

mixture within soot

formation limits (EQR 1,8

- 2,9).

Temperatures outside of

soot formation limits

(1350 – 1900 K) in rich

regions due to high EQR

ratios.

Interception of aromatic

species and limitation of

aromatic ring growth by

OH and HCO radicals.

Oxidation of carbon to CO

and CO2 within the flame,

reducing the amount of

carbon available for soot

precursor species.

Points of increased

reactivity where oxygen

groups are present,

reducing the threshold for

soot oxidation.

Developing of micropores

on the soot particles and

increase in active surface

for surface oxidation of C

to CO.

BA C

D




13

�̇� mass flow (kg/s)

𝜆 = 𝑚𝑎 (𝐿𝑠𝑡 · 𝑚𝑓)⁄ excess air ratio (-)

𝐿𝑠𝑡 stoichiometric air fuel ratio (-)

�̇� momentum flux (N)

𝑅 spray radius (m)

Based on equation (4), Figure 11 presents a cross section of spray cloud with indicated iso-lines for lower and

upper EQR limits of soot formation interval. As notable amount of oxygen is present in GLY, the lower limit of soot

formation (EQR = 2.9) is obtained relatively fast and close to the origin of the spray, whereas for D2, this limit is

positioned further away from the origin. The upper limit of soot formation (EQR = 1.8) for GLY is again closer to the

origin than for D2 due to its higher O/C ratio. The areas where soot formation is possible are indicated with red and

green shading. It is evident that the surface (which in 3D system corresponds to volume) of D2/air mixture within soot

formation interval is considerably larger, leading to more intense soot formation. This altered air concentration profile

thus influences not only the formation PM but also NOx emissions. Under the assumption that lower soot formation

limit roughly corresponds also to flammability limit of the fuel, the concentration of nitrogen is lower in the grey area

as less air is entrapped into the spray. The oxygen for lower flammability limit is mostly available already from the

fuel itself as was discussed above, so concentrations of nitrogen are lower in the green area when using GLY, whereas

in red area when using D2, the nitrogen to oxygen ratio is always 75/23.

Figure 11. Cross section of spray cloud with indicated EQR 2.9 and 1.8 for GLY and D2. Red area corresponds

to volume of D2/air mixture within soot formation limits. Green area corresponds to volume of GLY/air mixture

within soot formation limits.

3.5 Possible contaminants in crude glycerol, derived from biodiesel production

Albeit the technical difficulties linked to its viscosity and high oxygen content, the benefits of using GLY as a

fuel for microturbines are straight-forward when taking into account two of the most critical emission species, PM

and NOx. Although the presented results rely on technical grade glycerol and the application is still far from technically

perfected and commercial application, marketable benefit of GLY use for power generation would be the highest with

crude glycerol from biodiesel production, since its price per kWh is on par with natural gas. In this case, the limiting

parameter could be contaminants present in the crude glycerol which are a consequence of catalyst residuals and

carryover of feedstock’s mineral matter. Contaminant content significantly differs among different processes and

feedstocks. Base on available data analysis, compostion of crude glycerol, originating from different fatty acids in

Table 3 gives an insight into its technical feasibiltiy based on presence of contaminants. The data is taken from [4]

and comprises crude glycerol from 7 different feedstocks as indicated.




14

Table 3: Typical contaminants in crude glycerol from different feedstock and producers [4].

IdaGold Mustard

PacGold Mustard

Rapeseed Canola Soybean Crambe WVO

Calcium

[ppm] 11.7 +/- 2.9 23.0 +/- 1.0 24.0 +/- 1.7 19.7 +/- 1.5 11.0 +/- 0.0 163.3 +/- 11.6 BDL

Potassium [ppm]

BDL BDL BDL BDL BDL 216.7 +/- 15.3 BDL

Magnesium

[ppm] 3.9 +/- 1.0 6.6 +/- 0.4 4.0 +/- 0.3 5.4 +/- 0.4 6.8 +/- 0.2 126.7 +/- 5.8 0.4 +/- 0.0

Phosporus

[ppm] 25.3 +/- 1.2 48.0 +/- 0.2 65.0 +/- 2.0 58.7 +/- 6.8 53.0 +/- 4.6 136.7 +/- 57.7 12.0 +/- 1.5

Sulfur

[ppm] 21.0 +/- 2.9 16.0 +/- 1.4 21.0 +/- 1.0 14.0 +/- 1.5 BDL 128.0 +/- 7.6 19.0 +/- 1.8

Sodium

[wt.%] 1.17 +/- 0.15 1.23 +/- 0.12 1.06 +/- 0.07 1.07 +/- 0.12 1.20 +/- 0.10 1.10 +/- 0.1 1.4 +/- 0.16

Carbon

[wt.%] 24.0 +/- 0.00 24.3 +/- 0.58 25.3 +/- 0.58 26.3 +/- 0.58 26.0 +/- 1.0 24.0 +/- 0.00 37.7 +/- 0.58

Nitrogen

[wt.%] 0.04 +/- 0.02 0.04 +/- 0.01 0.05 +/- 0.01 0.05 +/- 0.01 0.04 +/- 0.03 0.06 +/- 0.02 0.12 +/- 0.01

*BDL – Below Detection Limit, *WVO – Waste vegetable oil

Listed contaminants are revealing the impact that transesterification process has on residual glycerol. Thus high

sodium content is a consequence of sodium methylate (NaOCH3) or sodium hydroxide (NaOH) used as a catalyst,

while Ca, K, Mg and P are a consequence of feedstock carryover which is linked to various parameters such as soil

conditions and process contamination with external substances. Considering the concentration, the most critical is Na,

which under conditions present in combustion chamber possibly reacts with SO2, CO2 and phosporus compounds,

leading to Na2SO4 (melting point 884°C), Na2CO3 (melting point 851°C) and possibly Na3PO4 (melting point 1583°C).

Hot path components in MGTs are particularly sensitive to ash particles with low melting points. These are at the

time they cross the turbine casing and rotor still in liquid phase (considering the operational conditions in Table 2).

Consequent slug formation, formed from molten ash usually sticks to the surface of hot components [38] which can

cause fouling problems that influence the efficiency and dynamics of the turbine rotor. Additional problem that occurs

with sulphates (such as Na2SO4) is the occurrence of hot corrosion through sulphidation attack. In some cases,

protective Cr2O3 scales on the surface of hot component material can dissolve in these sulphates [39] and remove the

protective layer, causing in-depth corrosion and removal of material underneath the damaged oxide layer. This

mechanism is active between temperatures 800°C to 950 °C which is also the most frequent temperature encountered

in MGT turbine rotors as well as heat exchangers. Thus, the operation of the MGT with crude glycerol could exhibit

at first reduced efficiency and then also reduced durability. On the other hand, ash particles with higher melting points

generally cause mostly abrasion damage which similarly influences the efficiency and durability of MGT and their

consequences are similarly challenging to tackle.

Apart from inorganic contaminants, crude glycerol contains also substantial amounts of methanol (in excess of

20%) and water (1-3 %). Although these components influence the combustion process, methanol is generally

elevating the calorific value and most negative impact can be attributed to water. The evaluation of impact of these

components require a dedicated research on glycerol-methanol-water blends which would as a result provide a suitable

primary measures to maintain the combustion process within reasonable emission limits.

3.6 Economic considerations of crude glycerol use for power generation

Based on market prices of crude glycerol and natural gas it is possible to estimate the business case of MGT

combined heat and power generation plant operating with these two fuels. This can be done with relatively simple

model for calculation of Return on Investment (ROI) interval, which provides the sensitivity of business case on key

parameters. The business case is valid under the assumption that technical challenges linked to contaminants are

resolved in the future.

The presented case takes into account a 100 kW MGT with the following estimated constraints for the use of

crude glycerol or natural gas:

Investment:

o 210.000 € (for MGT with adaptations required for crude glycerol utilization and installed on

pre-equipped location).

o 170.000€ (for MGT suitable for natural gas use, available from the shelf, installed on pre-

equipped location).

Maintenance:




15

o 1.600 €/year (for glycerol, including unexpected failures)

o 500 €/year (for natural gas)

Self-consumption:

o 20 kW (for glycerol with increased fuel system consumption, fuel manipulation, support

equipment)

o 10 kW (for natural gas)

Operational hours: 4000 h (yearly)

CO2 tax: 16 €/ton

Electricity selling price: 0.16 €/kWh (typical subsidiary price included – differs among countries)

Heat selling price: 0.04 €/kWh

Availability factor:

o 0,95 (for glycerol, considering the novelty of technology)

o 0,99 (for natural gas)

Fuel price:

o 0,04 €/kWh (for glycerol)

o 0,04 €/kWh (for natural gas)

Electric efficiency: 33 %

Thermal efficiency: 35 %

Taking into account upper estimated data, Figure 12 and Figure 13 present the sensitivity of ROI interval for natural

gas and crude glycerol, respectively with input parameters being listed above. Clearly, the ROI is notably lower with

natural gas (6,91 years), since it features higher availability, lower investment, lower CO2 emission and lower

estimated self-consumption. On the other hand, the ROI for crude glycerol operated MGT is 14.19 years due to higher

investment and self-consumption as well as lower predicted availability due expected reduction in maintenance

intervals, which notably impact the ROI interval. This accompanying expenses are also influencing the sensitivity of

crude glycerol operated MGT to fuel price, efficiency and number of operating hours, since the margin for

amortisation is lower.

Figure 12: Return on Investment interval for 100kW MGT operating on crude glycerol

Figure 13: Return on Investment interval for 100 kW MGT operating on natural gas




16

4. Conclusions

The main objective of this work was to obtain stable combustion and examine the combustion characteristics

of glycerol under regenerative cycle in MGT. This was achieved by implementing the key adaptations on the MGT

that enable the use of highly viscous preheated fuel.

While CO and THC emissions from GLY were significantly higher than for D2, it was observed that high

TIT and high primary air temperatures combined with high fuel preheating temperature could reduce the CO and THC

emission down to the levels, comparable with D2. This adjustments interfere very little with elevation of NOx

emissions and do not interfere at all with PM emissions as the NOx vs. PM,CO,THC trade off is not very pronounced

in case of GLY. The operational parameters of the MGT could therefore be adjusted to obtain diesel like CO and THC

emissions without penalty on NOx and PM. This clearly indicates the benefit of GLY as an alternative fuel since it

helps reduce the emissions below levels that are achievable with conventional diesel fuel.

Significant reduction of PM and NOx emissions when using GLY as a fuel was attributed to oxygen content

in the GLY. The elemental composition of GLY was reflected over the whole measurement range of the experimental

system. The reason for 10-fold reduction of PM emissions of GLY with comparison to D2 was attributed to six

underlying phenomena covered by three key mechanisms:

altered oxygen concentration profile during mixture preparation,

prevention of soot precursors formation and

promotion of soot oxidation reactions.

Altered oxygen profile during mixture preparation was also identified as the main mechanism for NOx

reduction. The identification of these mechanisms and their interrelation provide a basis for further improvement of

combustion process and exploitation of beneficial role of oxygen in viscous biofuels.

The importance of these findings also lies in the potential to increase the general acceptance of viscous waste

derived fuels which are often being linked to inefficient and environmentally questionable utilization. The results of

presented study oppose this view by obtaining diesel like levels of CO and THC emissions while at the same time

significantly reducing PM and NOx emissions and suggesting that highly viscous and highly oxygenated fuels might

be a suitable substitutes for decentralized power generation or even marine mobility by replacing heavy fuel oil.

To further support the transferability of results to crude glycerol from biodiesel production to industrial

turbines and heavy duty engines, an evaluation of possible contaminants revealed that sodium is the most critical as

its source lies in catalyst used for biodiesel production. This is a critical challenge which requires either purification

of crude glycerol which tends to reduce the economic attractiveness or further advances in fixed bed catalysts or

simply the use hot path components that are more ash resistant. At last, a short calculation of business case sensitivity

to various critical parameters was presented, revealing that ROI interval of MGT operating with crude glycerol would

exceed that of natural gas for roughly 8 years.

5. Acknowledgment

The authors acknowledge the support from Slovenian research agency through programme P2-0401 – Energy

engineering.




17

5. References

[1] Eggert H, Greaker M. Promoting Second Generation Biofuels: Does the First Generation Pave the Road?

Energies 2014;7:4430–45. doi:10.3390/en7074430.

[2] Directive 2009/28/EC of the European Parliament and of the Council of 23 April 2009 on the promotion of

the use of energy from renewable sources. http://eur-lex.europa.eu/legal-

content/EN/TXT/?qid=1421750499234&uri=CELEX:32009L0028; n.d.

[3] Ciriminna R, Pina C Della, Rossi M, Pagliaro M. Understanding the glycerol market. Eur J Lipid Sci

Technol 2014;116:1432–9. doi:10.1002/ejlt.201400229.

[4] Characterization of Crude Glycerol from Biodiesel Production from Multiple. Dep Biol Agric Eng

2006;3:2006–7. doi:10.1021/jf3008629.

[5] Seljak T, Rodman Oprešnik S, Kunaver M, Katrašnik T. Wood, liquefied in polyhydroxy alcohols as a fuel

for gas turbines. Appl Energy 2012;99:40–9.

[6] Seljak T, Oprešnik SR, Kunaver M, Katrašnik T. Effects of primary air temperature on emissions of a gas

turbine fired by liquefied spruce wood. Biomass and Bioenergy 2014;71:394–407.

doi:10.1016/j.biombioe.2014.09.016.

[7] Sallevelt JLHP, Gudde JEP, Pozarlik AK, Brem G. The impact of spray quality on the combustion of a

viscous biofuel in a micro gas turbine. Appl Energy 2014;132:575–85. doi:10.1016/j.apenergy.2014.07.030.

[8] Bohon MD, Metzger B a., Linak WP, King CJ, Roberts WL. Glycerol combustion and emissions. Proc

Combust Inst 2011;33:2717–24. doi:10.1016/j.proci.2010.06.154.

[9] Steinmetz S a., Herrington JS, Winterrowd CK, Roberts WL, Wendt JOL, Linak WP. Crude glycerol

combustion: Particulate, acrolein, and other volatile organic emissions. Proc Combust Inst 2013;34:2749–

57. doi:10.1016/j.proci.2012.07.050.

[10] Bohon MD, Roberts WL. NOx emissions from high swirl turbulent spray flames with highly oxygenated

fuels. Proc Combust Inst 2013;34:1705–12. doi:10.1016/j.proci.2012.07.064.

[11] Vera D, Jurado F, de Mena B, Schories G. Comparison between externally fired gas turbine and gasifier-gas

turbine system for the olive oil industry. Energy 2011;36:6720–30. doi:10.1016/j.energy.2011.10.036.

[12] Ensola GmbH. Technical Documentation Microturbine Turbec T100. Zurich 2006.

[13] Buffi M, Cappelletti A, Rizzo AM, Martelli F, Chiaramonti D. Combustion of fast pyrolysis bio-oil and

blends in a micro gas turbine. Biomass and Bioenergy 2018;115:174–85.

doi:10.1016/J.BIOMBIOE.2018.04.020.

[14] Seljak T, Oprešnik SR. Microturbine cCombustion and emission characteriszation of waste polymer-

derived fuels in a microturbine n.d.

[15] Seljak T, Kunaver M, Katrašnik T. Emission Evaluation of Different Types of Liquefied Wood. Strojniški

Vestn – J Mech Eng 2014;60:221–31. doi:10.5545/sv-jme.2013.1242.

[16] Seljak T, Katrašnik T. Designing the microturbine engine for waste-derived fuels. Waste Manag 2015;47:1–

12. doi:10.1016/j.wasman.2015.06.004.

[17] Seljak T, Širok B, Katrašnik T. Advanced fuels for gas turbines: Fuel system corrosion, hot path deposit

formation and emissions. Energy Convers Manag 2016:2016. doi:10.1016/j.enconman.2016.03.056.

[18] Seljak T, Pavalec K, Buffi M, Valera-Medina A, Chiaramonti D, Katrašnik T. Challenges and Solutions for

Utilization of Bioliquids in Microturbines. J Eng Gas Turbines Power 2018;141:031401.

doi:10.1115/1.4041312.

[19] Nayyar A, Sharma D, Soni SL, Bhardwaj B, Augustine M. Modeling and experimental investigation for

performance and emissions on a diesel engine using bio-oxygenated ternary fuel blends. Energy

2019;168:136–50. doi:10.1016/J.ENERGY.2018.11.098.

[20] Liu H, Wang Z, Long Y, Xiang S, Wang J, Wagnon SW. Methanol-gasoline Dual-fuel Spark Ignition

(DFSI) combustion with dual-injection for engine particle number (PN) reduction and fuel economy

improvement. Energy 2015;89:1010–7. doi:10.1016/J.ENERGY.2015.06.051.

[21] Zheng Z, Xia M, Liu H, Wang X, Yao M. Experimental study on combustion and emissions of dual fuel

RCCI mode fueled with biodiesel/n-butanol, biodiesel/2,5-dimethylfuran and biodiesel/ethanol. Energy

2018;148:824–38. doi:10.1016/J.ENERGY.2018.02.015.

[22] An H, Yang WM, Li J, Zhou DZ. Modeling study of oxygenated fuels on diesel combustion: Effects of

oxygen concentration, cetane number and C/H ratio. Energy Convers Manag 2015;90:261–71.

doi:10.1016/J.ENCONMAN.2014.11.031.

[23] SIST EN 590:2009/A101 Automotive fuels - Diesel - Requirements and Test Methods. n.d.

[24] Lefebvre AH, Ballal DR. Gas Turbine Combustion: Alternative Fuels and Emissions. vol. 54. CRC Press;

2011. doi:10.1002/1521-3773(20010316)40:6<9823::AID-ANIE9823>3.3.CO;2-C.

[25] Gupta KK, Rehman A, Sarviya RM. Bio-fuels for the gas turbine: A review. Renew Sustain Energy Rev

2010;14:2946–55. doi:10.1016/j.rser.2010.07.025.




18

[26] Chiaramonti D, Oasmaa A, Solantausta Y. Power generation using fast pyrolysis liquids from biomass.

Renew Sustain Energy Rev 2007;11:1056–86. doi:10.1016/j.rser.2005.07.008.

[27] Seljak T, Katrašnik T. DESIGNING THE MICROTURBINE ENGINE FOR WASTE-DERIVED FUELS

2014.

[28] Mansurov ZA. Soot Formation in Combustion Processes (Review). Combust Explos Shock Waves

2005;41:727–44. doi:10.1007/s10573-005-0083-2.

[29] Merker GP, Schwarz C. Gesamtprozeßanalyse. Tech. Verbrennung Simul. verbrennungsmotorischer

Prozesse, Wiesbaden: Vieweg+Teubner Verlag; 2001, p. 126–88. doi:10.1007/978-3-322-99823-1_3.

[30] Westbrook CK, Pitz WJ, Curran HJ. Chemical kinetic modeling study of the effects of oxygenated

hydrocarbons on soot emissions from diesel engines. J Phys Chem A 2006;110:6912–22. doi:Doi

10.1021/Jp056362g.

[31] Cheng AS, Dibble RW, Buchholz BA. The Effect of Oxygenates on Diesel Engine Particulate Matter, 2002.

doi:10.4271/2002-01-1705.

[32] Bechmann O. Untersuchungen zur Ablagerung von Rußpartikeln aus dem Abgas von Dieselmotoren 2000.

[33] Song J, Alam M, Boehman AL, Kim U. Examination of the oxidation behavior of biodiesel soot. Combust

Flame 2006;146:589–604. doi:10.1016/j.combustflame.2006.06.010.

[34] Seljak T, Rodman Oprešnik S, Katrašnik T. Microturbine combustion and emission characterisation of

waste polymer-derived fuels. Energy 2014;77:226–34. doi:10.1016/j.energy.2014.07.020.

[35] Cheng X, Chen L, Yan F, Dong S. Study on soot formation characteristics in the diesel combustion process

based on an improved detailed soot model. Energy Convers Manag 2013;75:1–10.

doi:10.1016/j.enconman.2013.05.033.

[36] Bechmann O. Untesuchungen zur Ablagerung von Russpartikeln aus dem Abgas von Dieselmotoren. 2010.

[37] Katrašnik T. Innovative 0D transient momentum based spray model for real-time simulations of CI engines.

Energy 2016;112:494–508. doi:10.1016/J.ENERGY.2016.06.101.

[38] Sreedharan SS, Tafti DK. Composition dependent model for the prediction of syngas ash deposition in

turbine gas hotpath. Int J Heat Fluid Flow 2011;32:201–11. doi:10.1016/j.ijheatfluidflow.2010.10.006.

[39] Lai GY. High-temperature corrosion and materials applications. ASM International; 2007.


seljak 2019 emission reduction through highly oxygenated

Documents