See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/281679945

Comparison of fuel value and combustion characteristics of two different RDF

samples

Article  in  Waste Management · September 2015

DOI: 10.1016/j.wasman.2015.08.037

CITATIONS READS

62 381

3 authors:

ayşe sever akdağ Aysel T Atimtay

Hacettepe University Middle East Technical University
3 PUBLICATIONS   93 CITATIONS    73 PUBLICATIONS   1,378 CITATIONS   

SEE PROFILE SEE PROFILE

Dilek F. Sanin
Middle East Technical University
81 PUBLICATIONS   2,121 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Change of microbial community composition in anaerobic digesters during the degradation of nonylphenol diethoxylate View project

All content following this page was uploaded by ayşe sever akdağ on 21 October 2021.

The user has requested enhancement of the downloaded file.

 Waste Management 47 (2016) 217–224

Contents lists available at ScienceDirect

Waste Management
journal homepage: www.elsevier.com/locate/wasman

Comparison of fuel value and combustion characteristics of two different

RDF samples
A. Sever Akdağ a,b, A. Atımtay a, F.D. Sanin a,⇑
a
Department of Environmental Engineering, Middle East Technical University, 06800 Ankara, Turkey
b
Department of Environmental Engineering, Hacettepe University, 06800 Ankara, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: Generation of Municipal Solid Waste (MSW) tends to increase with the growing population and economic
Received 12 May 2015 development of the society; therefore, establishing environmentally sustainable waste management
Revised 27 July 2015 strategies is crucial. In this sense, waste to energy strategies have come into prominence since they
Accepted 25 August 2015
increase the resource efficiency and replace the fossil fuels with renewable energy sources by enabling
Available online 7 September 2015
material and energy recovery instead of landfill disposal of the wastes. Refuse Derived Fuel (RDF),
which is an alternative fuel produced from energy-rich Municipal Solid Waste (MSW) materials diverted
Keywords:
from landfills, is one of the waste to energy strategies gaining more and more attention. This study
Combustion
Proximate analysis
aims to investigate the thermal characteristics and co-combustion efficiency of two RDF samples in
RDF Turkey. Proximate, ultimate and thermogravimetric analyses (TGA) were conducted on these samples.
TGA Furthermore, elemental compositions of ash from RDF samples were determined by X-Ray
Ultimate analysis Fluorescence (XRF) analysis. The RDF samples were combusted alone and co-combusted in mixtures with
coal and petroleum coke in a lab scale reactor at certain percentages on energy basis (3%, 5%, 10%, 20% and
30%) where co-combustion processes and efficiencies were investigated. It was found that the calorific
values of RDF samples on dry basis were close to that of coal and a little lower compared to petroleum
coke used in this study. Furthermore, the analysis indicated that when RDF in the mixture was higher
than 10%, the CO concentration in the flue gas increased and so the combustion efficiency decreased;
furthermore, the combustion characteristics changed from char combustion to volatile combustion.
However, RDF addition to the fuel mixtures decreased the SO2 emission and did not change the NOx pro-
files. Also, XRF analysis showed that the slagging and fouling potential of RDF combustion was a function
of RDF portion in fuel blend. When the RDF was combusted alone, the slagging and fouling indices of its
ash were found to be higher than the limit values producing slagging and fouling.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction has been utilized in the past few decades to solve both waste
and energy problems simultaneously (Rada and Andreottola,
Municipal Solid Waste (MSW) is an unavoidable by-product of 2012). RDF is the segregated high calorific fraction of processed
human activities. MSW generation increases with population MSW (Gallardo et al., 2014; Nasrullah et al., 2015; Rada and
growth and economic development as well as changes in lifestyles Ragazzi, 2014). The use of RDF in the thermal processes has
and in consumption patterns. Establishment of affordable, effective become popular and starts to receive wide attention in the world
and precisely sustainable Municipal Solid Waste Management as production of RDF provides a dramatic decrease in space
(MSWM) is crucial for sustainable development and promoting requirement and effectively utilizes the reusable energy of the
public health (Brunner and Rechberger, 2015; Cocarta et al., solid waste (Gug et al., 2015; Patel et al., 2012; Zhou et al.,
2009; Santibañez-Aguilar et al., 2013). 2013). Also, it is stated that its high energy content make RDF
As a sustainable approach, ‘waste to energy’ strategies are gain- compatible with conventional fossil fuels (Garg et al., 2007).
ing more attention. Production of Refuse Derived Fuel (RDF)/Solid Proximate and ultimate analyses are conducted in many studies
Recovered Fuel (SRF) is one of the waste to-energy strategies that to assess thermal characteristics of RDFs (Ahn et al., 2013; Dunnu
et al., 2010; Wagland et al., 2011). It is generally found that RDF
⇑ Corresponding author. samples have a low proportion of fixed carbon and a high amount
E-mail address: [email protected] (F.D. Sanin). of volatile matter when compared to other conventional fuels. In

http://dx.doi.org/10.1016/j.wasman.2015.08.037
0956-053X/Ó 2015 Elsevier Ltd. All rights reserved.
218 A. Sever Akdağ et al. / Waste Management 47 (2016) 217–224

addition, RDF samples contain higher moisture and have variable prepared by mixing RDF with the conventional fuels at ratios of 3%,
ash content (Ahn et al., 2013; Ferrer et al., 2005; Seo et al., 2010; 5%, 10%, 20% and 30%. The samples (alone and blends) were pre-
Wagland et al., 2011; Wang et al., 2014). Additional observations pared as pellets by a hand operated pellet press. The pellets had
on the combustion behavior of RDF are made by Thermogravimet- 15 mm diameter and about 2 mm height. Each sample (pellet)
ric Analysis (TGA). It is shown by Seo et al. (2010) that when RDF combusted was formulated to keep the calorific value constant as
amount in RDF-coal blends is higher, the rate of decomposition 4.184 kJ. Therefore, in this work, the RDF mixing ratio means the
increases and also the residual char decreases because of the high share of the energy input from the RDF in the total energy input
volatile fraction in RDF. Investigating the bottom ash composition from the fuel mixtures. RDF samples alone were also combusted
of RDF samples is also important in order to identify possible oper- without any coal or petroleum coke addition to observe the
ational problems such as corrosion, slagging and fouling. The high efficiency of RDF combustion alone.
presence of alkali metals in fuels is known to cause slagging,
fouling and ash agglomeration; this is especially a concern with
2.3. Thermal characterization of RDF, coal and petroleum coke samples
biomass originating fuels (Jenkins et al., 1998).
As mentioned above, there are a number of studies about RDF in
Proximate and ultimate analyses were made on the RDF, coal
literature; however, none of these studies make an integrated
and petroleum coke samples. Then the calorific values (lower heat
approach in evaluating the thermal properties of RDF by combining
value) of all the fuels were determined. Furthermore, XRF analyses
all the related features. In addition, studies evaluating the ash com-
were made on the ashes of the RDF samples used. All experiments
position by XRF and assessing this data for slagging and fouling
were done on triplicate samples.
potential of RDF are very limited. Therefore, the purpose of this
work is to investigate the thermal characteristics and combustion
efficiency of the RDF samples obtained from Material Recovery 2.3.1. Proximate analysis
Facilities (MRFs) of two big cities in Turkey by a holistic approach. In proximate analysis, it is typically assumed that the fuel was
In this approach, the thermal characteristics and combustion pat- composed of two types of carbon; volatile and fixed carbon. In this
terns of RDF samples is intended to be evaluated by proximate analysis, first, the moisture content of samples was determined
and ultimate analysis, TGA, and lab-scale combustion reactor according to ASTM method E790-08. Then, the volatile matter
studies. Additionally, comparative analysis with conventional fuels and ash content analyses were done in accordance with ASTM
is integrated into the study. The knowledge gap in literature on method E897-88 and ASTM method E830-87, respectively. Finally,
slagging and fouling potential of RDF is also expected to be filled the fixed carbon was calculated by subtracting the sum of percent-
by the results of this work. ages of volatile matter and ash content on dry basis from a total of
hundred percent.
2. Materials and methods
2.3.2. Ultimate analysis
2.1. The material recovery facilities investigated and sampling Carbon (C), oxygen (O), hydrogen (H), nitrogen (N) and sulfur
(S) are the main chemical elements in a fuel. The chemical analysis
The names of the cities that the RDF samples produced are not is very important to calculate the material balance accurately.
disclosed due to confidentiality reasons. So the two samples are Thus, C, H, N, S and O content of the samples were determined
denoted as RDF-A and RDF-B. In both facilities, MSW collected is by ultimate analysis conducted by Truspec Leco CHNS-932 ana-
processed in three lines. Fraction of waste that is paper, glass, plas- lyzer. Samples were introduced to auto sampler of the analyzer
tics and metal are recovered, the biodegradable fraction (mostly and combusted at 950 °C, and then the content of C, N, S and H
food waste) is sent to composting and the remaining portion is were measured simultaneously. Once the percentages of carbon,
processed as RDF. The MRFs schemes are given in Fig. 1. The RDFs nitrogen, hydrogen, sulfur and ash were determined, the amount
produced in both plants are mainly made up of fractions of textile, of oxygen was calculated by subtracting the total percentages of
paper and plastics. Major part of plastics is constituted by plastic the mentioned elements from a hundred percent. All results were
bags, with the remaining fraction PET. The composition of RDF reported on dry basis.
changes day to day but typical fractions are about 60% textile,
20% paper and 20% plastics. Since other organics including food
waste is diverted to a composting facility in both plants, all 2.3.3. Calorific values
remaining fraction exiting the RDF plant can be considered as com- Calorific values of RDF samples were determined according to
bustible. The RDF products are recently being combusted in differ- ASTM method E711-87 by using a Leco AC-500 model bomb
ent cement factories. calorimeter. The calorimeter was calibrated before each measure-
RDF samples, RDF-A and RDF-B, were taken on February, 2014 ment by using benzoic acid as a standard.
from the two material recovery facilities and were brought to the
laboratory for analysis. The RDFs were mixed well prior to sampling
2.3.4. X-Ray Fluorescence (XRF) analysis
in the facilities to obtain a representative sample. Upon arrival at
X-Ray Fluorescence (XRF) analysis was done by Spectro XLAB-
the laboratory, the RDF samples were ground well by the use of a
2000 PEDX-ray fluorescence spectrometer for the purpose of deter-
ball-mill to obtain homogenous samples and only the homogenized
mining the inorganic content of ashes of RDF samples. The device
samples were used throughout the study. Care was given to collect
was calibrated with the standard samples which are certified from
well-mixed samples before any analysis to further ensure the
U.S. Geological Survey (USGS). In the analysis, the loss of organics
homogeneity. The coal and petroleum coke samples used in the
on ignition was achieved by fluxing the sample with sodium
combustion studies were taken from a cement factory in Turkey.
tetraborate and burning at 1100 °C; then the inorganic content of
the sample remained was pelletized for further analysis. In this
2.2. Sample preparation for combustion analysis Na, Mg, Al, Si, P, S, Cl, K, Ca, Ti, V, Cr, Mn and Fe, which
were the major inorganic elements in the samples, were deter-
In order to observe the effects of RDF addition to main fuels dur- mined. Then, the elements were converted to their element oxides
ing combustion, RDF-coal and RDF-petroleum coke mixtures were to calculate the slagging/fouling indices of RDF samples.
 A. Sever Akdağ et al. / Waste Management 47 (2016) 217–224 219

Fig. 1. MRF schemes for RDF-A and RDF-B (left RDF-A, right RDF-B).

2.4. Thermogravimetric Analysis (TGA) of RDF samples analyzer consisted of two units which were a conditioning unit
and a gas analysis unit. In the conditioning unit called as PDG-
Thermogravimetric analysis was carried out according to ISO 100, moisture in the flue gas was trapped and solid particles in
11358-1 method by using Perkin Elmer Pyris STA- 6000 model the flue gas were filtered. Thus, the results of analysis were reported
TGA instrument, which measures the sample weight as a function on dry basis. In the second unit, the conditioned flue gas was ana-
of time as temperature is increased according to a determined lyzed. The device measured the concentrations of O2 and CO2 gases
heating rate. The instrument also calculates and displays Differen- with an Infrared (IR) cell as volume percentages; concentrations of
tial Thermal Gravimetric Analysis (DTGA) signal (rate of weight other gases were measured electrochemically in ppm.
loss) from weight loss vs. time signals. Using a constant heating
rate of 10 °C/min in dry air atmosphere, the samples were heated 3. Results and discussions
from 25 °C to 950 °C. TGA and DTGA curves for RDF-A, RDF-B, coal
and petroleum coke samples were obtained. 3.1. Thermal characterization of RDF, coal and petroleum coke samples

2.5. Lab-scale combustion experiments of RDF – coal and RDF - 3.1.1. Proximate/ultimate analysis of RDF, coal and petroleum coke
petroleum coke mixtures samples and their calorific value determination
The average results of proximate analysis and standard devia-
A laboratory scale set up was established in order to test the co- tions obtained from the triplicate analyses are given in Table 1.
combustion characteristics and efficiencies of RDF samples. In com- In evaluating the results, it should be noted that there is differ-
bustion experiments, pellets of RDF samples alone and in different ence in moisture and fixed carbon contents between RDF-A and
mixtures with coal and petroleum coke were combusted in an elec- RDF-B. This is because of the way the samples are processed in
trically heated cylindrical quartz batch reactor with dimensions of the facilities. The RDF-A sample was dried in its facility and then
50 mm diameter and 1200 mm height. A schematic diagram of sent to us, whereas RDF-B plant had no drying unit, so sample
the experimental setup is given in Fig. 2. The reactor had a perfo- was sent to us as it is and moisture content was determined by
rated plate made of quartz somewhere in the middle height of us as sampled. Besides, fixed carbon is the residue remaining after
the reactor to hold the sample pellets. The pellets were combusted volatile matter disappears during the combustion process. RDF
at the batch reactor where only one pellet was burned at a time at samples in this study were mainly composed of volatile matter
fixed temperature of 900 °C. The effect of RDF addition to the main (VM) which was 81.8% and 68.5% for RDF-A and RDF-B, respec-
fuels which were coal and petroleum coke was investigated by tively. As it is seen, the RDF-A has more volatile contents than
analyzing the flue gases during combustion experiments. RDF-B and thus its fixed carbon content is less. The reason of the
At the beginning of each experiment, the cylindrical quartz high volatile matter content of RDF is that they contain many vola-
reactor was heated first to 900 °C at which the combustion occurs. tiles such as paper, plastic etc.; also, the reduction of the inert
The two thermocouples, one of which located close to furnace wall materials from MSW stream during the RDF production process
and the other located inside the quartz reactor (combustion zone), increases the volatile matter content of RDFs.
allowed measuring the temperature of the furnace and the The volatile matter contents of RDF samples were higher than
combustion zone of the reactor. Then, air was introduced from that of the coal and petroleum coke samples, and their fixed carbon
the bottom of the reactor when the temperature of the combustion contents were lower accordingly. Among all samples, the petro-
zone reached 900 °C. The flow rate of air, which was 1000 cm3/min, leum coke sample had the lowest volatile matter content and high-
was adjusted by a flow meter. Next, top seal of the reactor was est fixed carbon content.
opened in order to drop pelletized sample into the reactor and In addition, the ash content of RDF is important since it deter-
was closed rapidly. The pelletized sample instantly reached the mines the calorific value and also the amount of residue that would
combustion zone of the reactor. The combustion began and the be left behind upon combustion. In this respect, RDF-A and RDF-B
gas analyzer started sampling the flue gas. were similar in ash content (between 12% and 15%) and ash con-
tents of both were slightly lower than that of coal (17.2%). On
2.6. Flue gas analyzer the other hand, petroleum coke had very small ash content
(0.7%) relative to the other three fuels.
The combustion gases, CO2, CO, NO, NO2 and SO2, were Carbon, hydrogen, and oxygen content of the fuels are very
monitored by Madur Photon flue gas analyzer continuously. The important since these elements constitute the main fuel fraction
220 A. Sever Akdağ et al. / Waste Management 47 (2016) 217–224

Fig. 2. Lab scale combustion setup.

Table 1
Proximate analysis of RDF samples, coal and petroleum coke.

Sample Moisture (%)* Volatile matter (%) **

Fixed carbon (%) **
Ash (%) **

RDF-A 1.6 ± 0.02 81.8 ± 0.89 5.2 ± 0.52 12.9 ± 0.38

RDF-B 14.8 ± 0.09 68.5 ± 0.52 16.6 ± 0.74 14.9 ± 0.64
Coal 4.3 ± 0.03 29.3 ± 0.11 53.5 ± 0.12 17.2 ± 0.22
Petroleum coke 7.0 ± 0.12 12.1 ± 0.17 87.4 ± 0.31 0.7 ± 0.01
*
% by weight, as collected.
**
% by weight on dry basis.

of the waste. Nitrogen content is also required to be known 3.1.2. Determination of ash composition of RDF samples by XRF
because it determines the amount of NOx formation. Sulfur is analysis
another important element, amount of which should be known The elemental composition of ash is important for slagging and
since the presence of sulfur causes the generation of acid gases fouling concerns. There is a wide range of elements which
SO2 and SO3 which contribute to air pollution and corrosion. contribute to slagging and fouling in combustion of refuse. XRF
In this context, ultimate analysis which gave the elemental analysis was done to determine the inorganic elements, which
composition of RDF, coal and petroleum coke samples were con- formed the ash content of RDF samples. Using the result of this
ducted and calorific values of the samples were determined all in analysis, fouling and slagging indices of RDF samples were calcu-
triplicate samples and the results with the standard deviations lated. Results of XRF analysis are presented as element oxides on
are shown in Table 2. dry basis in Table 3.
Ultimate analysis of the samples revealed that the carbon con- It can be seen in Table 3 that Si and Ca were the most abundant
tents of RDF-B, coal and petroleum coke were close to each other. inorganic elements in ashes of RDF samples tested in this study.
Carbon content of RDF-A was the lowest and that of the petroleum The Cl content in RDF ash samples was also high, which may orig-
coke was the highest among the samples studied. The calorific val- inate from chlorine in PVC (polyvinyl chloride) plastics and waste
ues of RDF samples, which were 22,143.40 kJ/kg and 19,229.79 kJ/ paper in RDF content (Beckmann and Ncube, 2007; Dong et al.,
kg on dry basis, for RDF-A and RDF-B, respectively, were close to 2002; Tchobanoglous and Kreith, 2002). Moreover, Fe, K, Al and S
each other and were a little lower than that of the coal used in this are other inorganic elements in ashes found in high concentrations.
study. The petroleum coke had the highest calorific value among Slagging and fouling indices were calculated for the RDF sam-
all the fuels due to its lower volatile matter content and higher ples by using the formula given in Table 4. Particularly ‘‘Base to
carbon content. Also, it was seen that the volatile matter content acid ratio” (B/A), and ‘‘sulfur ratio” (Rs) were used as slagging
decreases with an increase in the carbon content for all fuels. indices; ‘‘total alkalis” (TA), and ‘‘fouling index” (Fu) were used
Sulfur had the lowest percentage among all the elements as fouling indices, as they were previously used in Teixeira et al.
measured; in RDF-B the sulfur content was 0.45%, whereas sulfur (2012). B/A ratio shows the presence of alkali and alkali-earth oxi-
was not detected in RDF-A. Nitrogen and hydrogen percentages des in the sample relative to Al2O3, SiO2. TiO2 content have an
in the samples were all lower than 10%. However, the content of important role in melt formation. Higher basic compounds, B, are
hydrogen was higher in RDF samples than in coal and petroleum assumed to decrease the melting temperature, while the acidic
coke samples as seen in Table 2. The high hydrogen content of ones, A, increase the melting point. The sulfur ratio, Rs, is related
RDF samples may be linked to the high volatile matter content to slagging potential of sample and it is calculated using the S
originating from the organic matter in the waste. content of the sample. The fouling index, Fu, which is based on
 A. Sever Akdağ et al. / Waste Management 47 (2016) 217–224 221

Table 2
Ultimate analysis and calorific values of RDF samples, coal and petroleum coke.

Sample Carbon (%)* Hydrogen (%)* Nitrogen (%)* Sulfur (%)* Oxygen (%)* Calorific Value (kJ/kg)**
RDF-A 44.14 ± 1.09 5.63 ± 0.37 0.97 ± 0.31 – 36.26 ± 1.20 22,143.40 ± 661.1
RDF-B 56.47 ± 0.53 8.96 ± 0.25 1.50 ± 0.08 0.45 ± 0.02 17.73 ± 0.17 19,229.79 ± 151
Coal 63.80 ± 0.15 3.65 ± 0.02 1.88 ± 0.04 0.55 ± 0.02 12.94 ± 0.06 25,641.23 ± 279.1
Petroleum coke 68.12 ± 0.21 3.63 ± 0.05 1.94 ± 0.06 4.73 ± 0.09 20.51 ± 0.11 34,381.18 ± 299.2
*
% by weight on dry basis.
**
Lower heating value on dry basis (kJ/kg).

Table 3
Inorganic element contents of RDF samples (% by wt.).

RDF-A RDF-B
Na2O 0.20 0.06
MgO 0.13 0.12
Al2O3 0.43 0.42
SiO2 1.21 1.54
P2O5 0.14 0.11
SO3 0.20 0.24
Cl 0.29 0.32
K2O 0.27 0.44
CaO 3.36 2.72
TiO2 0.15 0.13
V2O5 – –
Cr2O3 – 0.01
MnO 0.01 0.01
Fig. 3. Results of co-combustion of coal with 5% RDF-A.
Fe2O3 0.08 0.23

concentrations in the flue gas with respect to time is shown on

Base-to-Acid ratio, gives more relevance to the alkaline elements the left ordinate of the graphs. The flue gas analyzer measured
which are the main agents of fouling. The results and limit values the concentration of these two gases as volume percentages. CO,
are given in Table 5. NO, NO2 and SO2 concentrations which were measured as ppm
As seen from Table 5, the RDF ash samples have moderate to are shown on the right ordinate of the graph. Measurements of flue
high slagging and fouling tendency when compared to limit values. gas composition were performed with two seconds interval; there-
Base to acid ratio (B/A) of both RDF ash samples were higher than fore, the profiles of gases were plotted with respect to time.
‘‘high slagging” class. Presence of alkaline elements in the RDF ash From the graph given in Fig. 3 and all other co-combustion
samples caused the high base to acid ratios, which were 2.3 and 1.8 experiments carried out in this study, it can be seen that a sharp
for RDF-A ash and RDF-B ash, respectively. It should be noted that decrease in oxygen concentration and increase in carbon monoxide
the values higher than 1 corresponds to high tendency of slagging. and carbon dioxide concentration were observed in the flue gas as
Fouling tendency of RDF ash samples was high when compared to soon as the pellets were dropped into the reactor. This quick car-
limit value because of the high content of volatile inorganic com- bon monoxide formation is explained by high volatiles content in
ponents like Na2O, K2O as shown in ash composition table. Also, RDF samples. Because the volatile matter burns quickly as soon
sulfur and phosphorus content of RDF samples are high, which as the pellet is dropped to the reactor, this creates an unstable
makes nearly all of slagging and fouling indices higher than the medium and results in high carbon monoxide concentration
limit values. It is worth to note that there are some studies in (Kobyashi et al., 2005). Another reason for high carbon monoxide
which slagging and fouling indices of coal is calculated (Teixeria concentration may be the temperature difference between the
et al., 2012; Park and Jang, 2011). As it can be seen from these stud- reactor and the fuel (pellet) at the beginning of the experiment.
ies, the slagging and fouling tendency of coal is lower when com- The pellet is at room temperature when it is dropped into the reac-
pared to RDF. tor. Incomplete combustion occurs until the pellet is heated up to
the reactor temperature, which results in high carbon monoxide
emission. Also, the fracture strength of the pellet has an indirect
3.2. Co-combustion of RDF–coal and RDF–petroleum coke mixtures effect on CO emission. The more brittle pellets are fractured into
small particles when they are dropped into the reactor and there-
Fig. 3 is given as a sample graph showing the results of a batch fore they burn faster. This fast burning increases the emission of
combustion experiment run. In this figure, change in O2 and CO2 volatiles and thus causes an unstable medium (Kobyashi et al.,
2005). On the average, oxygen level dropped to around 10% imme-
diately and carbon dioxide rose to around 5–10%. Concentration of
Table 4 carbon monoxide depended on the RDF percentage in mixtures; it
Empirical relations for slagging and fouling tendency of ash composition (Park and
changed between 1000 ppm and 5000 ppm when RDF percentage
Jang, 2011; Pronobis, 2005).
was 3–5%; on the other hand, it increased up to 6000 ppm when
Classification Index Formula RDF percentage was higher. The carbon monoxide concentration
Slagging Base-to-acid ratio (B/A) Fe2 O3 þCaOþMgOþK2 OþNa2 O increased to even higher values and it could not be measured with
SiO2 þTiO2 þAl2 O3
B

the gas analyzer for the mixtures with 20–30% RDF content or
Sulfur ratio (Rs) A xSðin dry fuelÞ
when sole RDF was burned. From the graphs, it can also be
Fouling Total Alkalis (TA) Na2 O þ K2 O

seen that the carbon monoxide formation was lower in
A xNa2 O þ K2 O
Fouling ratio (Fu) B
RDF-petroleum coke mixtures than the ones in RDF-coal mixtures.
222 A. Sever Akdağ et al. / Waste Management 47 (2016) 217–224

Table 5
Slagging and fouling indices calculated for RDF samples.

Slagging-fouling indices of RDF samples Limit values (Park and Jang, 2011; Pronobis, 2005)
RDF-A RDF-B Low Medium High
Slagging index B/A 2.3 1.8 <0.5 0.5 < B/A < 1 >1
RS 0.5 0.3 <0.6 0.6 < Rs < 2 >2
Fouling index TA 0.5 0.5 <0.3 0.3 < TA < 0.4 >0.4
Fu 1.1 0.9 60.6 – 0.6 < Fu 6 40

Petroleum coke, with higher calorific value and carbon content,

and lower volatile matter content than coal yielded better combus-
tion characteristics in RDF mixtures.
The efficiency of the combustion processes was calculated
based on a carbon mass balance. The percentage of carbon that
was captured as carbon monoxide and carbon dioxide was found
by calculating the areas under the emission curves for all experi-
ments. The carbon measured by the ultimate analysis was used
for the calculation of the amount of total carbon in the feed. These
calculations allowed us to check the degree of control that we had
on the system as well as the accuracy of our analysis during the
experiments. Calculations showed that carbon balance held within
a maximum of ±20% deviation, where in majority of the combus-
tions the deviation was much less than 10%. Then, the efficiencies
of the combustion were determined by calculating the amount of
carbon which transformed into carbon dioxide.
Fig. 5. Combustion Efficiency of RDF-B with coal and petroleum coke.
The results of carbon combustion efficiency for RDF-A and
RDF-B are shown in Figs. 4 and 5, respectively.
Both Figs. 4 and 5 show that the combustion efficiency
combustion yielding similar efficiencies. However, the test results
decreased with increasing RDF fraction. The replacement of coal
showed that the significant decreases were observed as the RDF
and petroleum coke with RDF caused more carbon monoxide for-
fraction increased from 10% to 20% in both coal and petroleum
mation which was the indication of incomplete combustion.
coke mixtures. Furthermore, the combustion efficiency also
ANOVA tests were performed to interpret the significance of the
decreased when the RDF ratio in petroleum coke mixtures was
results statistically. F-tests and multiple range tests were per-
increased from 20% to 30%. Also, the minimum efficiencies were
formed in the ANOVA test to compare the mean values of the
observed when only RDF samples were combusted without any
combustion efficiencies of RDF samples for the 7 different levels
coal and petroleum coke addition. Furthermore, for the co-
(0%, 3%, 5%, 10%, 20%, 30%, and 100%) of RDF fractions. It was found
combustion experiments with coal, it was seen that there was no
from F-test that there was a statistically significant difference
efficiency differences between the 20% and 30% mixing ratio. Also,
between the means of efficiency of RDF co-combustion with both
the general combustion efficiencies of petroleum coke–RDF mix-
petroleum coke and coal when different levels of RDF fractions
tures were higher than those of coal–RDF mixtures because of bet-
were incorporated into the fuel at 95.0% confidence level.
ter fuel features of petroleum coke compared to coal. It is
Multiple range test results indicated that there are no statisti-
concluded that the combustion efficiencies decreased significantly
cally significant differences between only coal or only petroleum
if the RDF sample fraction in the mixtures were more than 10%. At
coke combustion and co-combustion of coal or petroleum coke
this point it is worth to note that this percentage may change from
with the addition of 3%, 5% and 10% RDF. At these levels the
one condition to the other since the combustion efficiency highly
primary fuel (coal or petroleum coke) seems to dominate the
depends on the composition of RDF, which is variable even within
a single facility, combustion conditions such as temperature and
flowrate, reactor type and condition of samples (pellets or loose
material). However, the general trends are expected to be indepen-
dent of the working conditions. Therefore the results of this work
should be looked upon by considering these facts.
RDF addition to the fuel blends decreased the SO2 emission due
to the lower sulfur content in the RDF samples than in the coal and
petroleum coke samples, but it did not change NO emission appre-
ciably. SO2 emission was higher in petroleum coke mixtures than
coal mixtures since the sulfur content of petroleum coke was
higher than coal.

3.3. Thermogravimetric Analysis (TGA) of RDF, coal and petroleum

coke samples

TGA and DTG profiles evaluated in the combustion tests of RDF-

A and RDF-B and coal and petroleum coke are represented in Figs. 6
Fig. 4. Combustion Efficiency of RDF-A with coal and petroleum coke. and 7, respectively.
 A. Sever Akdağ et al. / Waste Management 47 (2016) 217–224 223

Fig. 6. TGA and DTG profiles of RDF-A and RDF-B (left RDF-A, right RDF-B).

Fig. 7. TGA and DTG profiles of coal and petroleum coke (left coal, right petroleum coke).

DTG curves of both RDF samples had shoulders/peaks which The combustion process in the experiments was complete at
were caused by the big mass losses observed in three different around 700 °C and the mass loss of RDF-A and RDF-B were found
regions. as 79.1%, 80.9%, respectively, which meant that about 20% of char
The first mass loss observed at the beginning the of the exper- remained at the end of the process.
iment, from ambient temperature to about 120 °C, is due to the Fig. 7 presents the TG and DTG curves observed for coal and pet-
loss of moisture and very light volatile matter content of the fuel. roleum coke samples. Unlike the RDF combustion, the main
In this region there is a difference between RDF-A and RDF-B that decomposition of coal samples took place at higher temperatures
matches nicely with the data in Table 1. The RDF-A has less mois- (between 700 °C and 800 °C). The highest mass loss, which corre-
ture, which yields a smaller peak at around 100 °C compared to sponded to the peak at 720°C, was mainly due to the volatile mat-
RDF-B. ter and char decomposition. At the end of the process the degraded
During the combustion process, the main decomposition of RDF mass was 53.3% of the original mass. The typical DTG curve char-
samples took place in between temperatures of 200 °C and 600 °C. acteristics for petroleum coke mentioned in Magdziarz and Werle
The rate of weight loss was maximum at around 320 °C for RDF-A (2014) were observed in the current study; there was only one
and at around 330 °C for RDF-B in the dry air environment. Unlike peak between the temperatures of 500 °C and 700 °C due to the
the moisture loss region, more than one peak was observed in this decomposition of all organic matter, loss of volatiles and char.
region. In the RDF-B curves, three exothermic peaks were observed The total weight loss in the combustion process was 35.8% and
at 330 °C, 400 °C, and 510 °C; while, in RDF-A curves, there were the maximum weight loss occurs at around 610 °C.
two peaks at 320 °C and 500 °C. At these peak points the mass loss The initial decomposition temperature of petroleum coke was
was significant. As mentioned in Piao et al. (2000), the presence of the highest when compared to coal and RDF samples. Therefore,
the shoulders shows that there are different volatile matter frac- having the highest ignition and burnout temperatures, petroleum
tions in RDF samples; for example the existence of cellulosic mate- coke was the hardest fuel to ignite and burnout totally among
rials and the plastics may produce different peaks during the samples burnt in the current study.
decomposition and combustion. When the decomposition temper- It was concluded from the thermogravimetric analysis that, the
atures and decomposition speeds of different types of volatile mat- constituents of the fuel decomposes in solid-phase at the char
ters included in RDF samples are close to each other, less peaks are combustion in coal and petroleum coke combustion; however, in
formed, which might be the case observed in RDF-A. RDF combustion, the decomposition occurs at early stages of the
For both RDF samples, a last peak occurred at about 650 °C. This combustion in gas phase due to the low fixed-carbon and high
last peak was due to the char combustion. Also the reactions volatile matter content of RDF. Also, it can be stated that the com-
between char and volatiles, which were coming from previous bustion mechanisms of volatile matters in RDF samples are compli-
phases of the process, might be another cause of this last peak. cated than that of char. Therefore, char combustion characteristics
224 A. Sever Akdağ et al. / Waste Management 47 (2016) 217–224

is not a criteria for RDF combustibility. However, volatile matter Cocarta, D.M., Rada, E.C., Ragazzi, M., Badea, A., Apostol, T., 2009. A contribution for
a correct vision of health impact from municipal solid waste treatments.
combustion is the most crucial parameter and should be investi-
Environ. Technol. 30 (9), 963–968.
gated since the combustion duration, speed and mass loss of vola- Dong, C., Jin, B., Zhong, Z., Lan, J., 2002. Tests on co-firing of municipal solid waste
tile matters affect the RDF combustibility. and coal in a circulating fluidized bed. Energy Convers. Manage. 43 (16), 2189–
2199. http://dx.doi.org/10.1016/S0196-8904(01)00157-1.
Dunnu, G., Maier, J., Scheffknecht, G., 2010. Ash fusibility and compositional data of
4. Conclusion solid recovered fuels. Fuel 89 (7), 1534–1540. http://dx.doi.org/10.1016/
j.fuel.2009.09.008.
Ferrer, E., Aho, M., Silvennoinen, J., Nurminen, R.-V., 2005. Fluidized bed combustion
Results of this study have shown that calorific values of RDF of refuse-derived fuel in presence of protective coal ash. Fuel Process. Technol.
samples on dry basis are close to that of coal and a little lower 87 (1), 33–44. http://dx.doi.org/10.1016/j.fuproc.2005.04.004.
compared to petroleum coke used in this study. Gallardo, A., Carlos, M., Bovea, M.D., Colomer, F.J., Albarrán, F., 2014. Analysis of
refuse-derived fuel from the municipal solid waste reject fraction and its
Although the calorific values of RDF samples are sufficient, the
compliance with quality standards. J. Clean. Prod. 83, 118–125.
use of sole RDF samples in combustion processes is limited due Garg, A., Smith, R., Hill, D., Simms, N., Pollard, S., 2007. Wastes as co-fuels: the policy
to its combustion characteristics. While the RDF sample fraction framework for solid recovered fuel (SRF) in Europe, with UK implications.
Environ. Sci. Technol. 41 (14), 4868–4874.
in the fuel mixtures is increased (such more than 10%), the com-
Gug, J., Cacciola, D., Sobkowicz, M.J., 2015. Processing and properties of a solid
bustion characteristics of the fuel mixtures changes from char energy fuel from municipal solid waste (MSW) and recycled plastics. Waste
combustion to volatile combustion. In addition, CO emission Manage. 35, 283–292.
increases and so the combustion efficiency decreases significantly ISO 11358-1, 2014. Plastics – Thermogravimetry (TG) of polymers – Part 1: General
principles.
when the RDF fraction in the mixture is higher than 10%. However, Jenkins, B., Baxter, Miles, T., 1998. Combustion properties of biomass. Fuel Process.
the co-combustion of RDF with lower ratios (such as 3%, 5%, and Technol. 54 (1–3), 17–46. http://dx.doi.org/10.1016/S0378-3820(97)00059-3.
10%) does not decrease the combustion efficiency significantly. Kobyashi, N., Itaya, Y., Piao, G., Mori, S., Kondo, M., Hamai, M., Yamaguchi, M., 2005.
The behavior of flue gas from RDF combustion in a fluidized bed. Powder
Also, the minimum efficiency is observed at the combustion of Technol. 151 (1–3), 87–95. http://dx.doi.org/10.1016/j.powtec.2004.11.038.
pure RDF without any coal or petroleum coke fraction. Magdziarz, A., Werle, S., 2014. Analysis of the combustion and pyrolysis of dried
RDF addition to the fuel blends decreases the SO2 emission due sewage sludge by TGA and MS. Waste Manage. 34 (1), 174–179. http://dx.doi.
org/10.1016/j.wasman.2013.10.033.
to the lower sulfur content in the RDF samples than in the coal and Nasrullah, M., Vainikka, P., Hannula, J., Hurme, M., Kärki, J., 2015. Mass, energy and
petroleum coke samples, but it does not change NO emission material balances of SRF production process. Part 3: solid recovered fuel
appreciably. produced from municipal solid waste. Waste Manage. Res. 33 (2), 146–156.
Park, S.W., Jang, C.H., 2011. Characteristics of carbonized sludge for co-combustion
Slagging and fouling intensity of RDF samples determined in
in pulverized coal power plants. Waste Manage. 31 (3), 523–529. http://dx.doi.
this study indicates that the use of RDF samples in co- org/10.1016/j.wasman.2010.10.009.
combustion processes needs to be kept in certain amounts. Patel, C., Lettieri, P., Germanà, A., 2012. Techno-economic performance analysis and
environmental impact assessment of small to medium scale SRF combustion
TGA results showed that the combustion mechanisms of vola-
plants for energy production in the UK. Process. Saf. Environ. 90 (3), 255–262.
tile matters in RDF samples are complicated than that of char. Piao, G., Aono, S., Kondoh, M., Yamazaki, R., 2000. Combustion test of refuse derived
The mass loss of RDF-A and RDF-B are found as 79.1%, 80.9%, fuel in a fluidized bed. Waste Manage. 20 (5–6), 443–447.
respectively, which means about 20% of char remained at the Pronobis, M., 2005. Evaluation of the influence of biomass co-combustion on boiler
furnace slagging by means of fusibility correlations. Biomass Bioenergy 28 (4),
end of the process. In coal and petroleum coke combustion, the 375–383. http://dx.doi.org/10.1016/j.biombioe.2004.11.003.
constituents of the fuel decomposes in solid-phase at the char Rada, E.C., Andreottola, G., 2012. RDF/SRF: which perspective for its future in the E.
combustion; however, in RDF combustion, the decomposition Waste Manage. 32 (6), 1059–1060.
Rada, E.C., Ragazzi, M., 2014. Selective collection as a pretreatment for indirect solid
occurs at early stages of the combustion in gas phase due to the recovered fuel generation. Waste Manage. 34 (2), 291–297.
low fixed-carbon and high volatile matter content of RDF. Santibañez-Aguilar, J.E., Ponce-Ortega, J.M., Betzabe González-Campos, J., Serna-
González, M., El-Halwagi, M.M., 2013. Optimal planning for the sustainable
utilization of municipal solid waste. Waste Manage. 33 (12), 2607–2622.
References Seo, M.W., Kim, S.D., Lee, S.H., Lee, J.G., 2010. Pyrolysis characteristics of coal and
RDF blends in non-isothermal and isothermal conditions. J. Anal. Appl. Pyrol. 88
Ahn, S.Y., Eom, S.Y., Rhie, Y.H., Sung, Y.M., Moon, C.E., Choi, G.M., Kim, D.J., 2013. (2), 160–167. http://dx.doi.org/10.1016/j.jaap.2010.03.010.
Application of refuse fuels in a direct carbon fuel cell system. Energy 51, 447– Tchobanoglous, G., Kreith, F., 2002. Handbook of Solid Waste Management. McGraw
456. http://dx.doi.org/10.1016/j.energy.2012.12.025. Hill Professional, U.S.A.
ASTM Standard E711-88, 2004. Standard Test Method for Gross Calorific Value of Teixeira, P., Lopes, H., Gulyurtlu, I., Lapa, N., Abelha, P., 2012. Evaluation of slagging and
Refuse-Derived Fuel by the Bomb Calorimeter. ASTM International, West fouling tendency during biomass co-firing with coal in a fluidized bed. Biomass
Conshohocken, PA. Bioenergy 39, 192–203. http://dx.doi.org/10.1016/j.biombioe.2012.01.010.
ASTM Standard E790-08, 2004. Standard Test Method for Residual Moisture in a Wagland, S.T., Kilgallon, P., Coveney, R., Garg, A., Smith, R., Longhurst, P.J., Simms, N.,
Refuse-Derived Fuel Analysis Sample. ASTM International, West Conshohocken, 2011. Comparison of coal/solid recovered fuel (SRF) with coal/refuse derived
PA. fuel (RDF) in a fluidized bed reactor. Waste Manage. 31 (6), 1176–1183. http://
ASTM Standard E830-87, 2004. Standard Test Method for Ash in the Analysis dx.doi.org/10.1016/j.wasman.2011.01.001.
Sample of Refuse-Derived Fuel. ASTM International, West Conshohocken, PA. Wang, G., Silva, R.B., Azevedo, J.L.T., Martins-Dias, S., Costa, M., 2014. Evaluation of
ASTM Standard E897-88, 2004. Standard Test Method for Volatile Matter in the the combustion behaviour and ash characteristics of biomass waste derived
Analysis Sample of Refuse-Derived Fuel. ASTM International, West fuels, pine and coal in a drop tube furnace. Energy Fuels 117, 809–824. http://
Conshohocken, PA. dx.doi.org/10.1016/j.fuel.2013.09.080.
Beckmann, M., Ncube, S., 2007. Characterisation of Refuse Derived Fuels (RDF) in Zhou, C., Zhang, Q., Arnold, L., Yang, W., Blasiak, W., 2013. A study of the pyrolysis
reference to the Fuel Technical Properties. In: 26th Annual International behaviors of pelletized recovered municipal solid waste fuels. Appl. Energy 107,
Conference on Incineration and Thermal Treatment Technologies, IT3; Phoenix, 173–182.
AZ; United States.
Brunner, P.H., Rechberger, H., 2015. Waste to energy – key element for sustainable
waste management. Waste Manage. 37, 3–12.

View publication stats