Waste Management 33 (2013) 2416–2424

Contents lists available at SciVerse ScienceDirect

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

Optimising energy recovery and use of chemicals, resources and

materials in modern waste-to-energy plants
J. De Greef a, K. Villani a, J. Goethals a, H. Van Belle a, J. Van Caneghem b,c,⇑, C. Vandecasteele b
a
Keppel Seghers, Center of Excellence, Hoofd 1, B-2830 Willebroek, Belgium
b
University of Leuven, Department of Chemical Engineering, ProcESS (Process Engineering for Sustainable Systems) Division, Willem De Croylaan 46, 3001 Leuven, Belgium
c
Group T Leuven Engineering College, Association of the University of Leuven, Andreas Vesaliusstraat 13, B-3000 Leuven, Belgium

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

Article history: Due to ongoing developments in the EU waste policy, Waste-to-Energy (WtE) plants are to be optimized
Received 19 February 2013 beyond current acceptance levels. In this paper, a non-exhaustive overview of advanced technical
Accepted 28 May 2013 improvements is presented and illustrated with facts and figures from state-of-the-art combustion plants
Available online 27 June 2013
for municipal solid waste (MSW). Some of the data included originate from regular WtE plant operation –
before and after optimisation – as well as from defined plant-scale research. Aspects of energy efficiency
Keywords: and (re-)use of chemicals, resources and materials are discussed and support, in light of best available
Waste-to-energy
techniques (BAT), the idea that WtE plant performance still can be improved significantly, without direct
SO2
SNCR
need for expensive techniques, tools or re-design. In first instance, diagnostic skills and a thorough under-
Boiler corrosion standing of processes and operations allow for reclaiming the silent optimisation potential.
Solidification Ó 2013 Elsevier Ltd. All rights reserved.
Stabilisation

1. Introduction
The thermal valorisation of municipal solid waste (MSW) is
subject to inherent limitations. Firstly, the physical appearance
and the chemical composition of MSW impede a cost-effective
and high-grade recovery of the lower heating value (LHV) con-
tained by the waste. Secondly, strict regulations with regard to
emissions and residue disposal tend to drive plant operators to-
wards steady operational margins. For example, pollution abate-
ment chemicals are often dosed in higher amounts than strictly
required, in order not to exceed the legal emission limit values
(ELV). Finally, where net profit generation is statutory not required
– i.e. when the treatment of MSW is directly undertaken by an offi-
cial authority or government Waste-to-Energy (WtE) plant opera-
tion is merely optimised to remain within acceptable budget,
even if more advantageous operation is technically feasible. In
the end, for a number of reasons, WtE plants are usually kept from
attaining high chemical and energetic efficiency levels.
In light of recent developments in the EU waste policy, it is ex-
pected that an increased number of plant operators will optimise
their processes of waste combustion, flue gas cleaning, heat recov-
⇑ Corresponding author at: University of Leuven, Department of Chemical
Engineering, ProcESS (Process Engineering for Sustainable Systems) Division,
Willem De Croylaan 46, 3001 Leuven, Belgium. Tel.: +32 16322353.
E-mail address: jo.vancaneghem@cit.kuleuven.be (J. Van Caneghem).
ery and energy production beyond current acceptance. In this pa-
per, experiences in the optimisation of state-of-the-art MSW
combustion are shared. Figures presented here originate either
from WtE plant data logging or from dedicated plant-scale research
initiatives. Such information appears to be lacking in scientific lit-
erature. A well-documented feedback from industrial WtE practice
is however important for further technological development of
WtE because many plant-scale aspects cannot be studied in a lab-
oratory environment. On the other hand, a waste combustion
plant’s operation is typically governed by economical drivers,
which is often not in favour of steady conditions required for
experimenting. Therefore it is not evident to perform WtE research
and dedicated process data analysis on the industrial scale.
The following topics are discussed:
 Impact of process stability on WtE plant performance.
 Consumption of NH3-based reagent for non-catalytic NOx
reduction.
 Early-stage assessment of boiler corrosion through chemical
analysis of boiler ash.
 Increase of heat recovery at the low-temperature end of WtE
boilers.
 Potential of residue recycling in dry lime (Ca(OH)2) based flue
gas scrubbing.
 Self-cementing behaviour of boiler/fly ash and APC residue in
solidification/stabilisation.

0956-053X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.wasman.2013.05.026
 J. De Greef et al. / Waste Management 33 (2013) 2416–2424
2. Issues in waste-to-energy plants
2.1. Process stability
Maintaining process stability is a challenge for combustion of
household waste. Firstly, the average physical appearance and
chemical composition of MSW change throughout the year. For
example, seasonal changes, and the occurrence of holiday periods
and weekends have a direct impact on both the amount and the
LHV of the waste as delivered to a WtE plant. Secondly, the inho-
mogeneous nature of MSW at any time causes instant disturbances
during the combustion process. As a result, the steam production
baseline (i.e. the average steam flow) of a WtE boiler exhibits a
fluctuating pattern over the longer term, with (superposed)
short-term fluctuations. Moreover, as for any other large industrial
system or process, the delaying behaviour of a boiler gives rise to
intrinsic over- and undershoots on the steam curve. Whereas the
latter is commonly tackled by means of conventional PID (Propor-
tional–Integral–Differential) control on the boiler as such, other
(waste-induced) effects are to be stabilised by the main control
system of the combustion. In this regard, it is worth referring to
a survey by El Asri and Baxter (2004), very concisely explaining
WtE plant control principles.
For a WtE plant in operation, the variation in the steam flow
from the boiler must be kept within a range acceptable for the
steam turbine. The highest electricity output is obviously obtained
when the steam production baseline is on average at the highest
level allowed for by design. However, when short-term fluctua-
tions are steadily beyond tolerance, the average steam production
(and hence the waste throughput) needs to be reduced to avoid
wear of the turbine. Good stability of the boiler steam produced
(and so of the combustion process) is hence required to maximise
the annual waste throughput and energy production.
2.2. Selective NonCatalytic Reduction (SNCR)
To date, WtE is already subjected to the lowest ELV amongst all
NOx emitting power industries within the EU (European parlia-
ment and council, 2010). Nevertheless, the ELV of 200 mg m3,1
is currently under political pressure and in a number of EU countries
(e.g. Sweden, Austria, the Netherlands, Flanders/Belgium, Bavaria/
Germany), more stringent emission limits are already applied. They
are either imposed explicitly by local law or established implicitly
(i.e. case-by-case) through environmental permitting or financial
incentives such as NOx tax/bonus policy.
NOx, i.e. the sum of NO and NO2, is a by-product of combustion.
In well-controlled combustion chambers, less NOx is formed, but
the formation of NOx cannot be totally avoided. It is therefore re-
quired to add an NH3-based reduction step to a MSW combustor
line in order to meet the daily ELV of 200 mg m3. Both Selective
Non-Catalytic Reduction (SNCR) as well as Selective Catalytic
Reduction (SCR) are effective and chemically well-understood
techniques. Due to the presence of a catalyst, SCR allows for a high
removal of NOx from the flue gas, even down to 30 mg m3. How-
ever, achieving such a low NOx emission is at the expense of a large
investment cost (i.e. about tenfold of SNCR) and a high CO2 equiv-
alent contribution. The latter is due to the production and mainte-
nance of the catalyst, the rather high minimum operation
temperature (>180 °C) and the input (or deviation) of energy/fuel
required. Given the overall limited energetic output of WtE plants
(compared to power plants), it was concluded by Beckmann et al.
(2009) that conventional SCR requires disproportionate efforts
1
m3’’ refers to 1 m3 at a temperature of 273.15 K, a pressure of 101.3 kPa, after
correction for the water vapour content and with an O2 content of 11%.
2417
per unit of NOx extra removed below 200 mg m3 and that therefore
SNCR is to be preferred if possible.
To date, a limited number of WtE-plants feature a high-dust SCR
system upstream of the flue gas cleaning. Such system is com-
monly positioned in or right behind the boiler and is operated at
higher temperature than a conventional SCR system. High-dust
SCR allows for a simpler and more compact NOx removal and is
energetically more advantageous compared to conventional SCR.
However, these plant references are all of recent date and typically
subsidised for purpose of research and development. First experi-
ences (in e.g. Brescia WtE plant, Italy) confirm high rate of clogging
and attrition of the catalytic units, as can be expected due to the
high dust load. Whereas catalysts in a conventional SCR system
can be used for more than 10 years, in a high-dust SCR system, a
partial catalyst replacement appears to be required already after
2 years of operation. Hence, maintenance costs are significantly
higher. Overall, high-dust SCR is still far from being a mainstream
application in the WtE industry, and more time is required to eval-
uate its techno-economical value waste combustion plants.
In new WtE plants a tendency exists to adopt conventional SCR
as soon as emission values are significantly lower than 200 mg m3
are to be established. In existing WtE plants then again, SNCR sys-
tems are often not well tuned to their technical capacity. To the
authors’ opinion, the intrinsic potential of SNCR technology for
application in WtE is not yet clear. Conditions of injection, mixing,
mass transfer (liquid–gas) as well as physical properties of the flue
gas in the zone of NOx reduction are critical for the effectiveness of
the NH3-reagent. Case studies by Villani et al. (2012) indicate that
there is still major room for improvement of SNCR systems in WtE
plants.
2.3. Corrosion assessment
Corrosion of boiler superheaters is well-recognized and causes a
significant maintenance cost in existing WtE plants. The key ele-
ments in the corrosion are identified as chlorine and sulphur,
which are present in MSW in higher doses than in other solid fuels.
Moreover, for MSW the Cl/S-ratio exceeds 2, i.e. the inverse of most
fossil fuels for power plants (Born, 2005). Cl and S together repre-
sent less than 1 wt% in MSW. Both elements form however highly
volatile compounds and are to a large extent transferred into the
flue gas during combustion at a temperature typically above
1000 °C, transfer coefficients for Cl and S being around 90% and
80% respectively (Belevi, 2000). As a result, concentrations in the
raw flue gas are typically 800–1200 mg m3 for HCl and 100–
300 mg m3 for SOx.
About15% of the SOx is rapidly captured during early cool-down
in solid salt deposits (on fly ash particles and boiler walls in the
post-combustion zone). The chlorine however remains for a longer
time in the flue gas, either as gaseous Cl2/HCl or as Na/K-chlorides,
that deposit at temperatures lower than 700 °C. As the gaseous Cl2/
HCl arrives at the superheater section downstream, corrosion is
triggered if the flue gas temperature still exceeds 650 °C and if
the superheater surface is at a contact temperature significantly
above 400 °C. If this situation persists for too long time, a liquid
phase containing alkaline Na/K-chlorides, will emerge on the cor-
roding surface. Finally, SO2 3 =SO4
2
ions also dissolve from ash
deposits and will act as electrolytic amplifiers, boosting up the
speed of corrosion by at least one order of magnitude. In general,
an oxidative chlorine-induced multiple reaction scheme is agreed
upon, although a number of unknowns is still remaining. For more
details on chemistry and mass transfer limitations, the reader is re-
ferred to e.g. Lee et al. (2007) and Horn et al. (2008).
On-line cleaning systems are helpful to restrict fouling on
superheater parts (in the convective boiler zone downstream),
but also on membrane walls (in the radiative boiler zone
2418 J. De Greef et al. / Waste Management 33 (2013) 2416–2424
upstream). By maintaining an adequate heat exchange during early
cool-down of the flue gas, the heat load and the temperature on the
superheaters downstream is limited and corrosion can be pre-
vented. Also inconel-type protection materials are widely used
for curing and preventing boilers from corrosion. However, the
cost-effectiveness of such materials on superheater surfaces in par-
ticular is disputed as the material composition was originally
developed for protection of evaporator surfaces (at lower temper-
ature). Furthermore, a single material is often applied in different
boilers with rather variable success as the method or procedure
of surface coating is critical (Herzog and Metschke, 2009). Hence,
new-to-build WtE boilers can benefit more from an evolved
know-how in thermal and geometric design, allowing for adequate
(residence) time, temperature (gradients) and turbulence of the
flue gas in different zones of the boiler (the ‘3T-theory’) (Albert,
1999). In this way, an optimal distribution of oxygen and temper-
ature is established, promoting sulphatisation of Na/K in the flue
gas. Alkaline sulphates deposit already in early stage in the boiler,
at high temperature, and thereby reduce the load of (detrimental)
alkaline chlorides in the flue gas downstream. A technological boi-
ler feature such as the Prism is proven supportive in this regard
(Adams and Périlleux, 2005).
For the purpose of early detection, analysis and prediction of
corrosion in WtE boilers, electro-chemical probes and sensors have
been developed, e.g. the heat flux sensor (Beckmann et al., 2007),
the Corrmorran probe (Waldmann et al., 2010) and the KEMCOP
(Zhan et al., 2010). However, a first-principles diagnosis might be
sufficient as a start to avoid costly equipment. Knowing that fly
ash particles are carried over from the combustion chamber and
are further formed on their track through the boiler, the composi-
tion of samples from consecutive ash sampling gates is a (partial)
reflection of the chemical events taking place. A comparison with
samples from other plants (with a known history) then yields early
indications on the occurrence of corrosion.
2.4. Low-temperature heat recovery
The thermal efficiency of a WtE plant can be improved, amongst
other things, by increasing the heat recovery from the flue gas. In
this regard, it is common practice to include external economizers
or tail-end heat exchangers near the lower temperature end of the
WtE process. Damage due to acid dew point corrosion is however
to be avoided when cooling down an unscrubbed flue gas. Hence,
in order to remove HCl and SOx (partially) before the flue gas
touches the low-temperature heat recovery units, the configura-
tion of the flue gas cleaning mostly ends up being complex and
expensive.
Still, lower limit temperatures of about 125 °C and 180 °C are
often imposed for boiler design, respectively on (incoming) boiler
feed water (BFW) and (outgoing) flue gas. Both plant-scale re-
search and operational experiences in several WtE plants however
indicate that dew point corrosion in MSW-originating flue gas is
triggered only far below these temperatures (Villani and De Greef,
2010). Obviously, the more heat is recovered from the flue gas in
the main boiler, the more straightforward the flue gas cleaning sys-
tem can be kept. In such cases, integrated rather than external
economizers cool the flue gas down to as low as 140 °C.
2.5. Lime and residue in flue gas cleaning
At temperatures below 160 °C, HCl and SOx can be well removed
from MSW-originating flue gas using dry Ca(OH)2-based scrubbers.
Due to the low operation temperature, this allows maximum heat
recovery before flue gas scrubbing (as briefly discussed in Sec-
tion 2.4). Dry Ca(OH)2 does not require any pre-handling on-site
(like e.g. preparing a lime milk) and hence offers the benefit of sim-
ple and clean operation. Moreover, lime is abundant and can be
sourced on a competitive market at reasonable cost. Also the resi-
dues (i.e. the mix of boiler/fly ash and lime reaction products) are,
after appropriate treatment, accepted on hazardous landfills
throughout the EU and abroad. These advantages turn dry lime-
based scrubbing into an overall attractive process.
On the other hand, dry flue gas scrubbers in WtE plants exhibit
slow reactions, as a low temperature has an adverse effect on reac-
tion and diffusion kinetics in a solid-gas system. In order to assure
compliance with imposed ELV’s, more Ca(OH)2 has therefore to be
added per unit of time, in overdose to the amount stoichiometri-
cally required. The stoichiometric ratio of lime dosing is typically
higher than e.g. for a semi–wet process involving a Ca(OH)2-
suspension or for a dry NaHCO3-based scrubbing. The overdosing
can however be strongly reduced by (partially) recycling the
residue retained in the bag filter, which still contains a significant
portion of unspent Ca(OH)2. Allowing it to react is then mainly a
matter of increased (residence) time in the flue gas cleaning
system.
The chemistry of reactions of Ca(OH)2 with HCl, HF and SOx are
in general well identified. From a historical viewpoint, desulfuriza-
tion of flue gas with lime(stone) at high temperature (>800 °C) has
been extensively studied for the needs of coal power industry. For
dry lime-based flue gas scrubbing in WtE plants, however, the is-
sue is rather to obtain dechlorination at low temperature
(<200 °C). As explained in Section 2.3, the conditions in MSW-
originating flue gas are complex and indeed largely differ from
those obtained with other fuels. Moreover, the effects of
competition between pollutant molecules on the kinetics of
Ca(OH)2 reactions and on and transport phenomena are not yet
well understood. Industrial flue gas cleaning practice is therefore
still highly empirical, so that systems in WtE industry are often
not operated at their intrinsic optimum.
2.6. Ash and residue treatment
Most of the boiler/fly ash and APC-residue of modern MSW
combustors does not meet the Waste Acceptance Criteria (WAC)
for hazardous waste landfills set by the European directive 1999/
31/EC and pursuant council decision 2003/33/EC. Primarily, the
leaching of chlorides (Cl), sulphates (SO24 ) and lead (Pb) is too
high (Amutha Rani et al., 2008). Hence, boiler/fly ash and APC-res-
idue have to undergo a dedicated treatment before final disposal
on a hazardous waste landfill. Solidification and stabilisation (S/
S) with Portland cement as a binder is worldwide the most used
treatment method (Quina et al., 2008; Chen et al., 2009). When
boiler/fly ash and/or APC-residue are mixed with Portland cement
and water is added, a hydration process is promoted, which is ex-
tremely complex, particularly due to the heavy metals present, of
which the fate and role in the different S/S steps in not yet fully
understood (Chen et al., 2009).
Advantages of using cement for the final treatment of a WtE’s
solid residues are that solidification and stabilisation are combined
in one step, and that the cost is relatively low. Disadvantages are
the increase of mass of the final product and the environmental
impact of the production of cement (mainly in terms of energy
use and CO2 emission). A large quantity of cement – typically
350 kg/t of residue – is to be added in order to comply with the
leaching standards (Dermatas and Meng, 2003; Quina et al.,
2008; Chen et al., 2009; European Commission, 2010). However,
as boiler/fly ash and APC-residues basically consist of the same
chemical compounds as Portland cement, i.e. CaO (40–50%), SiO2
(10–20%) and Al2O3 (5–10%), they might possess pozzolanic/ce-
ment like properties and so the question can be raised whether
the addition of cement is strictly required for the S/S of these
residues.
 J. De Greef et al. / Waste Management 33 (2013) 2416–2424
3. Results and discussion
3.1. Process stability
A typical example of a ten day’s operation in a WtE plant is
shown in Fig. 1a. Data curves of three selected parameters were ex-
tracted from the plant’s data logging system: waste throughput
(lower curve), primary combustion air (PA, middle curve) and
steam flow (upper curve). The waste throughput curve represents
the flow of waste fed into the combustor (Mg h1), in response to
the combustion control system. The flow (Mg h1) of steam
(400 °C at 40 bar pressure) can be considered as a measure for
the thermal energy effectively generated from the waste, since
the temperature and the pressure of the steam are set values given
a tight combustion control, with unaltered setpoint for steam pro-
duction, at a particular waste throughput, the PA flow (m3 h1)
then is regulated in reaction to LHV changes of the incoming waste.
Hence, fluctuations in the PA are a representative measure for (in-
stant) change of the calorific input of the combustor. As the effec-
tive steam output is separately tracked, it can be considered a
function of two independent variables i.e. the throughput and
the LHV of the waste.
In the initial situation (Fig. 1a), the waste throughput was
12.6 ± 1.2 Mg h1 on average, with about 29200 m3 h1 of PA, a
Fig. 1. Process data from a WtE combustor: waste throughput (lower curves, ton h1), primary air (PA) flow (middle curves, 103 N m3 h1) and steam production (upper
curves, ton h1). For a similar waste input as before (a), the steam setpoint could be increased by 1.4% after an optimisation of combustion control (b).
2419
steam flow of about 39.2 ± 1.7 Mg h1 and a calculated LHV of
9.6 ± 1.0 MJ kg1 (defined as l0 ± 1r0). Three months later, after a
PID-tuning of the combustion control system, data on the same
combustion parameters were logged again in the same plant
(Fig. 1b). For an identical average throughput of waste
(12.6 ± 0.7 Mg h1) with similar physico-chemical properties
(LHV 9.6 ± 0.5 MJ kg1), the overall stability of combustion had
clearly improved, resulting in a smoother feeding of the waste
and a lower consumption of PA (27100 m3 h1). Fluctuations in
steam production were now also reduced, i.e. from
r0  l1
0 6 4:4% down to r  l1 6 3:0% after optimisation. This
in turn allowed for raising the steam production set point with
1.2 Mg h1 (increase of 1.4%), resulting in 3.0% surplus of energy
produced, calculated as ðl  l0 Þ  l1
0 . Moreover, from the reduc-
tion in PA (7.5%), a 20% saving of electric power on the side of
the PA air fans can be estimated (for air fans: power is proportional
to flow3).
3.2. SNCR
A situation with deficient use of ammonia reductant in the WtE
boiler is illustrated in Fig. 2 (left side). About 45 kg h1 of the agent
was injected with only moderate success. Although the NOx emis-
sion was below the ELV of 200 mg m3, the slip of unconsumed
2420 J. De Greef et al. / Waste Management 33 (2013) 2416–2424
ammonia at the stack was on average above 10 mg m3. The ammo-
nia was injected with two lances only at one side of the first empty
boiler pass (in the post-combustion zone). When the ammonia
dosage was diminished, the NOx emission almost immediately in-
creased. Despite the relatively small size of the boiler (27 MWth), it
was clear that the coverage of the boiler’s cross-section obtained
by the SNCR was sub-optimal.
Two injection lances were added on the opposite wall at the
same height to achieve a better distribution of the ammonia, and
the SNCR system was tuned to the new situation. It can be seen
from Fig. 2 (right side) that finally the ammonia slip was reduced
to well below 10 mg m3. To obtain a similar NOx emission
(180 mg m3), the ammonia consumption was now 20% less.
Clearly, the distribution (and the uptake) of ammonia in the flue
gas had been limiting the chemical efficiency of the SNCR.
3.3. Corrosion assessment
Two horizontal-type boilers in different WtE plants are com-
pared. In both installations, mixed MSW is combusted with LHV’s
of about 9.5 MJ kg1 and 7.5 MJ kg1, respectively. The waste input
of both plants is nevertheless considered comparable for the actual
purpose, as no significant difference appears in the composition of
the inorganic fraction of the waste (figures not included here). Me-
tal elements and chlorine in the flue gas originate predominantly
from the inorganic fraction in the waste (Niessen, 1995). The dif-
ference in LHV is due to a larger fraction of organic waste and inert
materials in the MSW fed to the second boiler. While the high
water content in organic waste causes an evaporation enthalpy
loss during combustion, the inert materials obviously release nei-
ther heat nor chemical elements. Notwithstanding the resulting
difference in (adiabatic) furnace temperature between the two
plants (1150 °C and 1000 °C, respectively), volatilisation tempera-
tures are exceeded for all considered chemical elements, such that
partitioning equilibriums can be regarded well established in both
cases.
All ash ports below the vertical and horizontal passes were
sampled on the two different WtE-boilers. The chemical analysis
Fig. 2. Impact of SNCR remediation: ammonia slip (lower curve, right Y-axis) and consumption (middle curve, right Y-axis) for a similar NOx emission (upper curve, left X-
axis). The difference was established by adding two injection lances. While the unreacted ammonia is now reduced to about 5 mg m3, the overall consumption of ammonia is
20% less.
results of all ash samples, expressed in wt% on dry basis, are pre-
sented in Fig. 3 (a: first plant; b: second plant). In the second plant
fewer samples could be collected, due to less sampling ports being
available or accessible. The results in the graphs shown in Fig. 3 are
however aligned according to similarity in sampling location (x-
axis).
It appears that the chemical composition of ash samples for
similar positions strongly differs between the two boilers. The first
boiler (a) is designed and maintained such that heat exchange pat-
terns and temperature profiles are established as was expected by
design. Hence, a conventional pattern in ash composition is ob-
tained including 50–75% Si-, Ca-, Mg- and Al-oxides for all samples.
The collection of chlorides and sulphates is about similar along the
entire length of the boiler (610%), with exception of the third
superheater (‘‘EVA1 + SH3’’) position, where this concentration in-
creases to 25%. Since this is the hottest boiler part, some fouling
commonly occurs by desublimation of elements from the flue gas
directly onto the superheater surface. The slightly higher content
of Na, K, chloride and sulphate at the position of ‘‘EVA1 + SH3’’
can then be explained by bits of the fouling, removed from the
superheater surface by the cleaning system, that are collected to-
gether with boiler ash particles, dropping out from the flue gas
stream. A value of 25% for the sum of chloride and sulphate is a
rather high value already, but is perhaps due to little reserve in
thermal design of the boiler. In this plant, it has been noticed that
the flue gas temperature indeed locally approaches 650 °C, i.e.
about the upper limit for a corrosion-free operation. Nevertheless,
a common distribution of the temperature and chemical elements
is still obtained throughout the boiler and no major corrosion
occurs.
In the second boiler (b) on the contrary, the overall amount of
inorganic oxides is very low, i.e. on average <50%. At ‘‘top pass
1’’, the deposit on the boiler walls contains little Na and K
(<10%), whereas in the superheater zone these elements are found
in very high amounts (up to more than 80%). It indicates an ineffec-
tive sulphatisation of alkalines (explained in Section 2.3): too little
in the post-combustion zone and too much downstream. Nonethe-
less, given the relatively high organic fraction in the waste, sulphur
 J. De Greef et al. / Waste Management 33 (2013) 2416–2424 2421

100% 100%

80% 80%

60% 60%

40%
40%

20%
20%
(a) (b)
0%
0%
3
ht
ft
1

A2

1
2

1
H
le

rig
ss

O
SH

SH
+S

3
1
EV
3

A1

1
EC

EC

EC

1
2

H
Pa

2+

O
SH

SH
ss
A1

+S
ss

ss
2+

EV

EC

EC

EC
p

pa
ss

pa

pa
EV

A2
To

ss
Pa

EV
Pa

to
SiO2 CaO MgO Al2O3 TiO2 Cr2O3 Fe2O3 MnO P2O5 Na2O SiO2 CaO MgO Al2O3 TiO2 Cr2O3 Fe2O3 MnO P2O5 Na2O
K2O SO3 CuO NiO V2O5 PbO ZnO Cl Sr+Ba K2O SO3 CuO NiO V2O5 PbO ZnO Cl Sr+Ba

Fig. 3. Chemical composition of boiler ash sampled in two different WtE plants (a and b). The sampling position is indicated on the horizontal axis. (‘‘Pass 1’’, ‘‘Pass 2’’ and
‘‘Pass 3’’ are vertical boiler passes; ‘‘SH’’ = superheater; ‘‘EVA’’ = evaporator; ‘‘ECO’’ = economiser).

should be sufficiently available in the flue gas (Niessen, 1995), and 107%
B
is therefore not considered the limiting factor for sulphatisation. 106%
At the time of sampling, the flue gas around the superheaters 105%
104% A
was however found to have a temperature of 650–700 °C, whereas Relative 103%
in the lab a sintering temperature below 600 °C was measured for Electrical 102%
the corresponding ash samples (‘‘Pass 3’’, ‘‘EVA2 + SH3’’ and Efficiency 101%
100%
‘‘SH2’’). It is known that fly ash particles with a relative composi-
99%
tion as in Fig. 3b are still sticky under these temperature conditions 98%
and thus contribute to a rapid build-up of fouling on the super- 97%
heater tubes. Severe corrosion was indeed also reported after some 96%
time. It became clear that a combination of inadequate thermal de- 135

145
125
120

175
sign and insufficient boiler cleaning obstructed the heat exchange 115 FG temp

210
BFW temp (°C) 105
between flue gas and water/steam in the boiler. (°C)

3.4. Low-temperature heat recovery
Simulations were made with CycleTempoÒ modeling software
(developed by TU Delft, the Netherlands) on a simple Rankine cycle
containing boiler, turbine, condenser, condensate preheating and
de-aerators as main parts. Input parameters for the thermody-
namic model were taken from heat and mass balance calculations
for a steam boiler (52 bar, 400 °C superheating) fired with waste of
13 MJ kg1. The exercise was repeated several times combining dif-
ferent boiler feed water (BFW) inlet and flue gas outlet tempera-
tures. To establish a clear comparison, all steam is converted into
electricity such that the electricity output at the turbine is a
straightforward indicator for the steam flow produced by the boi-
ler. In Fig. 4 the electricity yield is presented on a comparative
scale. Other process and steam cycle parameters were kept identi-
cal for the different cases.
The reference case (100%) is a typical combination of tempera-
tures one would encounter in older WtE-plants with a semi-dry
flue gas cleaning (210 °C FG vs. 135 °C BFW). When lowering the
flue gas temperature (movement direction ‘‘A’’ in Fig. 4) to a typical
combination for a bicarbonate system (at 175 °C), a relative in-
crease of electrical efficiency of more than 3% can be achieved. A
dry lime-based system (at 145 °C) even allows for 6% of efficiency
increase. This is however a theoretical result, since a DT of 10 °C
between boiler feed water and flue gas would require an unaccept-
ably expensive boiler design with a very large economiser section.
Fig. 4. Relative (gross) electricity yield for different combinations of boiler feed
water and flue gas temperatures at the exit of a WtE boiler. A strong reduction of
the design flue gas temperature (A) requires an (economical) correction in feed
water temperature (B). Nevertheless, a significant efficiency gain of >4% can be
obtained by adapting the outlet temperature conditions.
A downward correction in BFW temperature (movement direction
‘‘B’’ in Fig. 4) is necessary to restore a reasonable economiser size.
Although a part of the (theoretical) gain has to be given up, a sig-
nificant gain in the order of 5.5% (125 °C BFW) and 4% (105 °C
BFW) is still possible when extending the lower temperature sec-
tion of the boiler. This result illustrates that a compact boiler, fea-
turing reduced design temperatures for the outgoing flue gas as
well as the incoming boiler feed water, can be an energetically
interesting option for WtE plants.
3.5. Lime and residue in flue gas cleaning
The flue gas cleaning system of Grundon’s medical waste incin-
erator in Lakeside (UK), where a test campaign was performed,
consists of two consecutive stages (Fig. 5). Each stage comprises
a tower with venturi section and a fabric filter unit. Hydrated lime
is injected at the bottom of the tower, which acts as an entrained
plug flow reactor. Reacted lime and fly ash particles are then
trapped on filter bags and regularly purged off as residue. This
mixture is collected in the buffer under the bag filter and then
2422 J. De Greef et al. / Waste Management 33 (2013) 2416–2424

re-injected as recyclate in the tower reactor. The flow of recyclate

can be adapted; a purged amount is permanently removed from
the system as residue for disposal.
For modest pollutant levels, the double-staged plant configura-
tion allows for operation on the second stage while experiments
are done on the first stage, without the risk of exceeding the ELV’s.
Inlet values of HCl and SOx are measured upstream of the first stage
and emissions are monitored with an accredited continuous emis-
sion monitoring system (CEMS) at the stack. For the purpose of the
test, a portable CEMS was temporarily installed between the two
stages of the system, in order to observe the first stage only. Sam-
ples were also taken regularly from the recyclate buffer and sent to
a lab for chemical analysis (Fig. 6).
Within the scope of this paper only the study of recyclate buffer
capacity is highlighted, although different features of the system
were studied. The curves in Fig. 7 below are produced from inlet
and outlet emission values for HCl and SOx of the first stage. The
removal percentage plotted, is calculated as the concentration dif- Fig. 6. Content of unreacted lime (Ca(OH)2), Cl and S in samples of the recycled
residue.
ference (out minus in), divided by the inlet concentration, at a gi-
ven time. Before the test, it was ensured that the buffer was
replenished (dosing of fresh Ca(OH)2 at stoichiometric ratio of
100%
2.5 for about one day). For the test itself, the supply of fresh lime
was ceased and the recycling was set to full speed, leaving the 98%
recyclate buffer saturating for 9 h. During this period, no excep- 96%
tional inlet pollutant values occurred, i.e. the inlet concentrations
94%
varied within a range (defined as l ± 1r) of 997 ± 365 mg m3 for
HCl and 204 ± 120 mg m3 for SOx. The test was repeated two days 92%
removal

later, with slightly different inlet conditions: 1080 ± 470 mg m3 for 90%
HCl and 270 ± 156 mg m3 for SOx. In between both tests, fresh lime
88%
was added with a stoichiometric ratio of about 1. Hence, the recyc-
late buffer was not replenished and kept (on average) in a condi- 86%
tion where it was left at the end of the first test. The second test 84%
was therefore considered as the continuation of the first, with re-
82% HCl
gard to unreacted Ca(OH)2 content of the recyclate. Intermediate
residue analysis justified this approach (Fig. 6). 80%
0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00
Starting with a fresh buffer (test 1), analysis of Fig. 7 shows that
hours
the removal of HCl remains almost at maximum (99–97%), while
SOx removal drops from 95% down to about 80%, i.e. only 1.4% loss Buffer Test 1 Buffer Test 2

in removal per hour. When starting off with the non-replenished

buffer (test 2), the removal of both HCl and SOx decreases much 100%
more rapidly, as expected. At the end of the second test, the re- 90%
moval of SOx had dropped to about 15%. The SOx emission at the
stack never exceeded the ELV of 50 mg/m3, because SOx was fur- 80%
ther removed in the second stage of the system. 70%
60%
removal

3.6. Ash and residue treatment 50%

40%
Boiler ash and (CaO-based) APC residue were sampled from a
MSW-fired WtE plant, where S/S is currently accomplished on-site 30%
20%
Ca(OH)2 Activated
Carbon 10% SO2
0%
0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00
Flue Gas hours
OUT
Buffer Test 1 Buffer Test 2

Fig. 7. Potential of a residue buffer in a dry Ca(OH)2-based flue gas cleaning system.
Fresh lime supply to the system was stopped and only residue was internally
recycled for up to 9 h. The removal efficiency of HCl remains high (90%). The
Recycling Recycling removal efficiency of SOx drops more rapidly, but still ends up at 30% after a long
Flue Gas & Buffer & Buffer depletion.
IN

Fig. 5. A double-staged, dry Ca(OH)2-based flue gas cleaning system (Grundon’s

medical waste incinerator, Lakeside, UK). Due to back-up capacity in the second
by use of cement. I.e. per Mg of blend (30 wt% boiler ash + 70 wt%
stage, experiments could be performed to map the system behaviour of an APC residue) 350 kg of cement is dry mixed and after the addition
industrial-scale flue gas cleaning unit including residue recycling. of water, a granular, solidified and stabilised product is obtained,
 J. De Greef et al. / Waste Management 33 (2013) 2416–2424 2423

Table 1
Average leaching of heavy metals from boiler ash, APC residue and their S/S mixtures (mg kgDM1).

Boiler ash APC-residue S/S water onlya S/S cementb WAC haz. wastec WAC non-haz. wastec
pH 12.8 12.6 11.6 12.8 – –
SO2 25,354 26,013 2,307 257 50,000 20,000
4
As 0.86 3.85 1.1 0.86 25 2
Ba 7.0 85 10 7.7 300 100
Cd 0.07 0.09 0.04 0.02 5 1
Cr 17 5.0 5.0 3.6 70 10
Mo 5.4 3.5 <5 <5 30 10
Ni 0.51 1.6 0.33 0.44 40 10
Pb 366 1,465 2.2 14.5 50 10
Sb 0.05 0.10 0.28 <0.01 5 0.7
Se 3.42 0.47 1.1 0.68 7 0.5
Zn 59 101 30 23 200 50
a
Average values of all different boiler ash/APC-residue mixtures tested.
b
30% boiler ash-70% APC residue mixture on-site stabilised with cement.
c
Waste acceptance criteria for hazardous and non-hazardous waste landfills set by European Council Decision 2003/33/EC.

which is landfilled. It was interesting to explore whether boiler ash
and APC-residue can be sufficiently stabilised without the use of
cement.
The boiler ash and APC residue samples were mixed in varying
proportions (0/100, 10/90, . . ., 90/10, 100/0). To 20 g of each mix-
ture ultra pure water was added until a sticky paste was obtained.
The amounts of water added ranged from 4.5 ml for the mixtures
containing at least 70% boiler ash to 8.5 ml for the pure APC-resi-
due sample. The samples were then left to cure for one month.
To simulate (worse case) environmental conditions in the landfill,
the cured solidified monolithic samples were ground with a pestle
in a mortar until granules were smaller than 4 mm. Leaching of
heavy metals and anions was determined in one step leaching
tests: 100 ml of ultra pure water was mixed with10 g of sample
and shaken for 24 h. The leachate was filtered using a 0.45 lm
membrane filter and, if necessary, diluted with ultra pure water
before analysis. For heavy metals ICP-MS (Thermo-Elemental X-
series) was used, for anions ion chromatography (DIONEX IC
2000). For comparison untreated boiler ash and APC residue and
a sample of a (30% boiler ash 70% APC residue) mixture on-site sta-
bilised with cement were submitted to the same leaching test. The
results in Table 1 show that untreated boiler ash and APC residue
exceed the WAC for hazardous waste landfills for Pb and cannot
be landfilled as such. The leaching of Zn and SO2 4 is high, but below
the EU WAC limits for hazardous waste landfills. The leaching of Cr,
Ni, As, Se, Mo, Cd, Sb and Ba is well below the EU WAC limits. The
leachate pH is high as explained by the presence of Ca, Mg, K and
Na oxides. Whereas for the untreated ash and residue samples the
measured Pb concentrations exceed the legal limits, Table 1 shows
that they can be reduced well below the WAC for hazardous waste
landfills and even below the WAC for non hazardous waste land-
fills by S/S with water only. Also the average leaching concentra-
tions of other regulated heavy metals and SO2 4 can be brought
below the WAC for non-hazardous waste landfills, with exception
of Se. Table 1 gives the average leaching values for the different
boiler ash APC residue mixtures. Table S1 in the supplementary
material gives the results for the individual mixtures, showing that
Pb leaching highly depends on the pH. Table S1 (supplementary
material) shows that Pb leaching only exceeds the WAC (50 mg/
kgDM), when the leachate pH is 12.5 or higher. In the cement stabi-
lized boiler ash/APC mixture, the leaching of Pb remains below the
50 mg kgDM1 limit, although the pH of the leachate was 12.8 (Ta-
ble 1). Obviously, due to the presence of cement in the mixture,
other mechanisms than hydroxide formation apply for Pb. The
pH dependence of the Pb leaching from the mixtures S/S by adding
only water was confirmed in a second series of tests (Table S2 in
supplementary material). It was however also observed that the
pH of the leachate decreased to below 12.5 and hence the leaching
of Pb to below 50 mg kgDM1 when the samples were left to cure
for 12 more days (Table S2 in supplementary material). This is
due to the natural carbonatation i.e. reaction of the basic compo-
nents in the hardened samples e.g. CaO with CO2 from the air. Ta-
ble 1 also shows that for the other heavy metals leaching from the
samples S/S by adding only water is in the same order of magni-
tude than from the sample stabilised with cement. For sulphates,
the leaching was about 10 times higher for the samples S/S by add-
ing only water than for the cement-stabilised sample, but re-
mained well below the WAC for non hazardous waste landfills
These results indicate that S/S of boiler ash and APC-residue, by
adding water only, is a valid, yet ecologic and economic alternative
for conventional cement based S/S.
4. Conclusions
Since its introduction, waste combustion has developed from
poorly designed incineration to a thermal treatment technology
meeting high technical, sanitary and environmental standards.
Furthermore, modern WtE has proven to be the only reliable and
cost-effective technique up to now for thermal treatment of large
amounts of mixed MSW. Therefore it has become an indispensable
part of sustainable waste management in many developed
countries.
However, due to a number of (non-technical) reasons, opera-
tional and design practice in WtE is not everywhere at optimum
yet. Based on selected experiences in the optimisation of combus-
tion, energy recovery and production, pollutant abatement and
residue treatment, it was illustrated that results can be obtained
with the existing technology, beyond what is generally considered
achievable. In was also demonstrated that a basic diagnostic ap-
proach, involving the required knowledge, can also lead to a thor-
ough understanding of a process or a problem, so that expensive
tools and techniques can be avoided where possible.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.wasman.2013.
05.026.
References
Adams, B., Périlleux, M., 2005. Design features to avoid high-temperature corrosion
in MSW boilers. In: Saxonia (Ed.), Dampferzeugerkorrosion – Teil 1. Wagner
Verlag, Großschirma, pp. 247–260.
2424 J. De Greef et al. / Waste Management 33 (2013) 2416–2424
Albert, F., 1999. Korrosion bei müllgefeuerten Dampferzeugern: technisch
wirtschaftliche Überlegungen, In: VDI-Berichte 1484. VDI-Verlag, Düsseldorf,
pp. 1–34.
Amutha Rani, D., Boccaccini, A., Deegan, D., Cheeseman, C., 2008. Air pollution
control residues from waste incineration: current UK situation and assessment
of alternative technologies. Waste Manage. 28, 2279–2292.
Beckmann, M., Gampe, U., Grahl, S., Hellfritsch, S., 2009. Zielkonflikt emissionen und
energieeffizienz am beispiel der 37.BImSchV. In: Thomé-Kozmiensky, K.J.,
Beckmann, M. (Eds.), Energie aus Abfall, vol. 6. TK Verlag, Neuruppin, pp. 3–32.
Beckmann, M., Krüger, S., Magel, G., Spiegel, W., 2007. Online-bestimmung des
wärmestroms auf membranverdampferwände von dampferzeugern. In:
Thomé-Kozmiensky, K.J., Beckmann, M. (Eds.), Energie aus Abfall, vol. 2. TK
Verlag, Neuruppin, pp. 641–658.
Belevi, H., 2000. Factors that control the quality of municipal solid waste
incineration residues. In: Dhir, R.K., Dyer, T.D., Paine, K.A. (Eds.), Sustainable
construction: use of incinerator ash. Thomas Telford Publishing Ltd., London,
pp. 19–31.
Born, M., 2005. Wissenschaftliche grundlagen der rauchgasseitigen dampferzeuger-
korrosion. In: Saxonia (Ed.), Dampferzeugerkorrosion – Teil 1. Wagner Verlag,
Großschirma, p. 33.
Chen, Q., Tyrer, M., Hills, C., Yang, X., Carey, P., 2009. Immobilization of heavy metal
in cement-based solidification/stabilization: a review. Waste Manage. 29, 390–
403.
Dermatas, D., Meng, X., 2003. Utilization of fly ash for stabilization/solidification of
heavy metal contaminated soils. Eng. Geol. 70, 377–394.
El Asri, R., Baxter, D., 2004. Process control in municipal solid waste incinerators –
survey and assessment. Waste Manage. Res. 22, 177–185.
European Commission (2010). Reference document on best available techniques in
the cement, lime and magnesium oxide manufacturing industries.
Herzog, T., Metschke, J., 2009. Cladding(ge)schichten – erfahrungen als grundlage
für qualitätsanforderungen. In: Thomé-Kozmiensky, K.J., Beckmann, M. (Eds.),
Energie aus Abfall, vol. 6. TK Verlag, Neuruppin, pp. 505–516.
Horn, S., Haider, F., Waldmann, B., Warnecke, R., 2008. Quantifizierte betrachtung
des stoffübergangs in belägen und korrosionsschichten zur beschreibung der
korrosions-geschwindigkeit, In: Beläge und Korrosion, Verfahrenstechnik in
Großfeuerungsanlagen. VDI-seminar 430508, Oberhausen.
European parliament and council, 2010. Directive 2010/75/EU on industrial
emissions (integrated pollution prevention and control).
Lee, S.H., Themelis, N.J., Castaldi, M.J., 2007. High-temperature corrosion in waste-
to-energy boilers. J. Therm. Spray Technol. 16, 104–110.
Niessen, W., 1995. Combustion and incineration processes – applications in
environmental engineering, second ed. Marcel Dekker, New York.
Quina, M., Bordado, J., Quinta-Ferreira, R., 2008. Treatment and use of air pollution
control residues from MSW incineration: an overview. Waste Manage. 28,
2097–2121.
Villani, K., De Greef, J., 2010. Exploiting the low-temperature end of WtE-boilers, In:
Proceedings of the third Intl Symposium on Energy from Biomass and Waste,
Venice.
Villani, K., De Greef, J., Goethals, J., de Montauban, I., Van Langenhove H., 2012.
Exploring the performance limits of non-catalytic de-NOx in waste-to-energy
plants, In: Proceedings of the fourth Intl Symposium on Energy from Biomass
and Waste, Venice.
Waldmann, B., Schrupp-Heidelberger, D., Stöcker, B., Haider, F., Horn, S., Maisch, S.,
Warnecke, R., Müller, V., 2010. Online-korrosionsüberwachung als instrument
zur bekämpfung der dampferzeugerkorrosion – strategien verwirklichen und in
echtzeit überprüfen. In: Thomé-Kozmiensky, K.J., Beckmann, M. (Eds.), Energie
aus Abfall, vol. 7. TK Verlag, Neuruppin, pp. 289–300.
Zhan, H., de Jong, M., Veldman, H., Lamain, J., 2010. Large scale material testing and
on-line corrosion monitoring at the high efficiency WtE-plant of AEB.
Presentation PowerGen 2010, Amsterdam.