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

net/publication/324824893

Optimization Of The Dry Flue Gas Efficiency Loss Of Boiler Dsanilation Plant
Using RSM

Article · April 2017

DOI: 10.37376/1571-000-030-010

CITATIONS READS
0 886

2 authors:

Naji Abdelwanis Omer M. Elmabrouk

Clemson University University of Benghazi
9 PUBLICATIONS   17 CITATIONS    10 PUBLICATIONS   4 CITATIONS   

SEE PROFILE SEE PROFILE

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

THE USE OF SIMULATION FOR PROCESS IMPROVEMENT IN BENGHAZI CANCER TREATMENT CENTER View project

Using Fuzzy Logic in Prediction View project

All content following this page was uploaded by Naji Abdelwanis on 28 April 2018.

The user has requested enhancement of the downloaded file.

 OPTIMIZATION OF THE DRY FLUE GAS
EFFICIENCY LOSS OF BOILER DSANILATION
PLANT USING RSM

Naji Abdelwanis1 Omar M Elmabrouk2

1
Department of Mechanical Engineering-Omar Al-Mukhtar University
2
Department of Industrial & Manufacturing Systems Engineering- Benghazi University
1
[email protected]
2
[email protected]

ABSTRACT In this study an attempt was used confidently in optimizing operating

made to develop a response surface model to parameters that would minimize LDG
optimize loss due dry flue gas (LDG) of Key words: loss due dry flue gas (LDG),
boiler efficiency at Sussa’s desalination optimization, response surface method
plant. The most significant operating (RSM).
parameters studied were combustion air flow
rate (MC), mass flow rate of fuel (MF) and 1. Introduction
flue gas temperature (TF), and their Direct and indirect approaches are well
corresponding ranges were (80.3-91.5 known used to calculate boiler efficiency as
Kg/hr.), (6-6.5 Kg/hr.), and (309.1-374.6 °C) following:
respectively. The MINITAP software 1- Direct method can be calculated by using
version 16 was used to design the the output (steam) and the heat input (i.e.
experiments (DOE), to mathematically fuel). Equation 1 is used to calculate the
evaluate the effect of the operating efficiency using direct method [1].
𝐻𝑒𝑎𝑡 𝑂𝑢𝑡𝑝𝑢𝑡
parameters on the LDG and to set the 𝐵𝑜𝑖𝑙𝑒𝑟 𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 𝜂 = × 100 (1)
𝐻𝑒𝑎𝑡 𝑖𝑛𝑝𝑢𝑡
optimal operating parameters that provide a Figure 1 shows the schematic diagram of the
minimum value of the response (LDG). operating parameters affecting the boiler
NORSOK M-506 software was used to efficiency
calculate LDG for each experiment. The
best response value was analyzed using the
response surface and contour plots. The
optimal operating parameters were 80.3kg/hr
for combustion air flow rate, 6.50 Kg/hr for
mass flow rate of fuel, and 309.10 for flue
gas temperature, and their corresponding
LDG value was 4.317.4 kJ°C /kg . It could Fig.1. Boiler efficiency operating parameters [2]
2- Indirect method can also be used to
be concluded that historical-data RSM is a
calculate the boiler efficiency using equation
promising statistical technique that could be
number 2. [1]
𝜂 = 100 − (𝐿1 + 𝐿2 + 𝐿3 + 𝐿4 + 𝐿5 + 𝐿6 ) (1)
Where: thus maximize heat transfer that leads to
L1- Loss due to dry flue gas (LDG) increase the efficiency of the boiler at
L2- Loss due to hydrogen in fuel Sussa’s desalination plant.
L3- Loss due to moisture in fuel
L4- Loss due to moisture in air 2. Literature Review
L5- Loss due to carbon monoxide Evaporator (where desalination is made) and
L6- Loss due to surface radiation, boiler are the most important components of
convection and other unaccounted. any desalination plant. In Sussa desalination
Figure 2 shows the schematic diagram of the plant there are two evaporators with capacity
losing parameters affecting the boiler of 20000T/day, a flow rate of 93.2T/hr at a P
efficiency 25 bar and T of 224o C, the maximum
temperature is at cell 1 which is 64.1o C,
each one consists of 5 cells cell 1,2,3 are
divided in two vessels (1A-B , 2 A-B , 3A-B )
where A and B are operates at the same
conditions , cells 4,5 are simple in
construction. There are two boilers model
1SM01A1 and 2SM01A1 in Sussa’s
desalination plant each rated at 126 T/hr. of
Fig.2 Losses parameters affecting boiler efficiency[2] superheated steam at a pressure of 24 bars.
Each boiler consumes about 6.3 T/hr. of
Indirect method (Losses Method) is more heavy fuel oil (HFO).
accurate than direct method because it has As known the boiler efficiency reflects how
too many error sources [2]. In this study, the the boiler will transfer the heat. It has a great
losses method has been used for boiler impact on heating related energy.
efficiency calculations. Based on the previous equations boiler
Dry flue gas losses, radiation losses and efficiency can be improved by two methods.
other losses can significantly affecting the The first one is by optimizing boiler’s inputs
boiler efficiency [3]. Loss due dry flue gas such as fuel and air inputs, which is more
loss (LDG) is considered to be the major challengeable. In this approach key variables
losses affecting the boiler efficiency [2, 4]. must be identified and adjusted in order to
Equation 3 is used to calculate the LDG as maximize the boiler efficiency. The other
following: ( method is by minimizing the boiler losses.
)
mc ×C pg × Tf −Ta
L1 = × 100 (3) Several studies have been conducted to
Mf ×C v
Where, improve boiler efficiency in different
Mc: combustion air flow rate (Kg/hr). industrial sectors by minimizing the heat
Mf: mass flow rate of fuel (Kg/hr). losses [2,4,5]. Minimizing heat losses such
Cp : specific heat of flue gas (KJ/Kg.°C). as air to fuel ratio is really challenges [6].
Tf: flue gas temperature (o C). Therefore, it is important to find an
Ta: ambient temperature (o C). appropriate technique to minimize such
CV: calorific value of fuel (KJ/Kg).i these losses and, thus optimize the boiler
In this paper, the response surface method efficiency.
was used to investigate the effect of In any boiler the main heat loss results from
operating process parameters, namely discharging hot gases into the chimney [7].
combustion air flow rate (MC), mass flow Heat can be lost from boilers by a variety of
rate of fuel (MF) and flue gas temperature ways such as dry flue gas losses, moisture in
(TF), on the boiler efficiency and to fuel losses, moisture from burning hydrogen
determine the optimal values of these losses, moisture in air losses, unburnt losses
parameters that minimize the LDG, and the in fly ash as carbon, sensible heat in bottom
ash losses, radiation and convection losses,
formation of carbon monoxide and blow the boiler water and carrying it up the stack
down losses[Gupta, Ghai and Jain [8], Moni [5, 7]
Kuntal Bora and Nakkeeran [2]. As Generally, the excess air depends on fuel
mentioned above, dry flue gas loss is type and amount of oxygen [7]
considered as the greatest boiler loss [2, 10]. Gupta, Ghai and Jain [8] suggested several
In the worst case losses could be reach up to recommendations to improve the overall
30%−35% of the boiler efficiency Gupta, efficiency such as controlling excess air and
Ghai and Jain addressed that [8]. Bora and employing better insulation on steam pipes.
Nakkeeran has stated that, out of the 23% Shieh, Chang, Jang, Mac and Huang [4]
total losses, there were 11.36 % losses due used a stepwise regression to identify the
to heat loss due to dry flue gas[2]. key variables of the thermal efficiency
Saidur, Ahamed and Masjuki have described estimator of multi-fuel boilers. The thermal
that the major amount of energy lost efficiency was improved by 1.94 and 0.73%
through flue gases is due the fact that the of a boiler burning coal and multi-gas in the
produced heat cannot transferred to water or virtual plant simulations.
steam inside the boiler. In addition, they Using Indirect method technique to
addressed that the temperature of the flue calculate the boiler efficiency Harish and
gas leaving the boiler ranges from 150 to Tej Pratap[5] have stated that the maximum
250o C, about 10-30% of the heat is lost efficiency of boiler can be obtained for the
through it [3]. Therefore, it is important to range of 20-40% of excess air and the boiler
evaluate the energy outputs and efficiency to maximum combustion zone can be found
analyze the different heat losses in order to between 3.5 to 5% of O 2 with a flue gas
identify the main causes of these losses that temperature between 134o C and 135o C.
lead to reduce the boiler efficiency. Finding Therefore, using indirect method will
these factors will lead to find the appropriate provide the optimum LDG that will increase
method to recover them and, thus to improve the efficiency of the Sussa’s desalination
the overall boiler efficiency. A substantial plant.
save in energy will be reached, if such these
losses can be reduced [7]. To achieve the 2. ANOVA (Analysis Of variance)
high boiler efficiency stack losses, dry flue For this paper, the response surface
gases loss and water vapor (moisture) loss methodology (RSM) is used to investigate
should be minimized. the effect of the operating parameters
Dry flue gas losses depend on amount of selected on the response (LDG).
excess air and the exit temperature of flue Central composite design (CCD) is one of
gas[8], and typically 4% or greater. the primary design techniques in RSM. This
Excess air has been consistently shown to technique is commonly used for process
be significantly related to the heat loss and it optimization.
could reach up to 140% of excess air for The three main parameters affecting the
poorly maintained boilers Einstein, Dan; LDG of the boiler efficiency at Sussa’s city
Worrell, Khrushch and Marta[10 ]. This is desalination plant located at the east cost of
happen because the air enters the boiler with Libya are shown in table 1. Table 1 shows
an ambient temperature and leaves with high the operating parameters and their
temperature, taking a large amount of heat corresponding ranges
with it [8]. The more excess air, the more
exhaust gas, the more extra flue gas loss. On
the other hand, if the air for combustion is
insufficient it will lead to wastage of flue
because of incomplete combustion, and thus
reduce the heat transfer efficiency absorbing
heat that would otherwise be transferred to
 Table 1 The operating parameters and LDG investigated to determine more accurately
ranges the optimum boiler efficiency operating
parameters.
Range In this study, ANOVA table is made with
Parameters Notation Unit help of MINITAB 16 software. When
Lower Upper
value value performance varies one determines the
Combustion average loss by statistically averaging the
MC Kg/hr 80.3 91.5 quadratic loss. The average loss is
air flow rate
Flue gas proportional to the mean squared error of Y
MF Kg/hr 6 6.5 about its target T. The initial techniques of
temperature
flue gas o the analysis of variance were developed by
TF C 309.1 374.6
temperature the statistician and geneticist R. A. Fisher in
the 1920s and 1930s, and are sometimes
The purpose of the statistical analysis of known as Fisher's ANOVA or Fisher's
variance (ANOVA) is to investigate which analysis of variance, due to the use of
design parameter significantly affects the Fisher's F distribution as part of the test of
LDG. Based on the ANOVA, the relative statistical significance.
importance of the boiler efficiency
parameters (with respect to LDG) is

Table 2 ANOVA table 1

Source DF Seq SS Adj SS Adj MS F P
Regression 9 5.01092 5.01092 0.55677 267210.30 0.000
Linear 3 4.99348 4.99348 1.66449 798841.34 0.000
MF 1 0.46397 0.46397 0.46397 222673.98 0.000
MC 1 1.23131 1.23131 1.23131 590941.93 0.000
TF 1 3.29820 3.29820 3.29820 1582908.11 0.000
Square 3 0.00039 0.00039 0.00013 61.97 0.000
2
MF 1 0.00039 0.00021 0.00021 98.44 0.000
2
MC 1 0.00000 0.00000 0.00000 0.02 0.879
TF2 1 0.00000 0.00000 0.00000 0.02 0.879
Interaction 3 0.01705 0.01705 0.00568 2727.60 0.000
*
MF MC 1 0.00157 0.00157 0.00157 752.53 0.000
MF* TF 1 0.00423 0.00423 0.00423 2031.06 0.000
*
MC TF 1 0.01125 0.01125 0.01125 5399.21 0.000
Residual Error 10 0.00002 0.00002 0.00000
Lack-of-Fit 5 0.00002 0.00002 0.00000 * *
Pure Error 5 0.00000 0.00000 0.00000
Total 19 5.01094
0.05 indicate the model terms are not
Table 2 shows the Analysis of Variance significant.
(ANOVA) for LDG response. The important
information can be obtained here is the The regression parameters of the developed
percentage influence of all factors over model of the response with statistical
responses. P value less than 0.05 indicate significance were calculated the main
model terms are significant. In this case MF, interactive relationship between the
MC, TF, MF*MC, MF*TF and MC*TF are experimental parameters and response were
significant model terms. Values greater than
evaluated. The response of the experiment is 3.Responce Surface Modeling
represented as the form of equation (1) LDG Once the response , parameters and levels
have been selected , the next step is to
design .
After the parameters and the values are input
into the software (MINITAP16), a DOE
LDG= 5.37415 – (0.21540mf) (1) model will be automatically generated with
+(0.35090mc)+(0.57430Tf) specific number of runs coupled with
+( 0.00864mf2 )+ (0.00014)mc2 ) – specific parametric settings. In this case
(0.00014Tf2 )- (0.01400mf*mc) – twenty runs were generated as shown in
(0.02300mf*Tf) +(0.03750mc*Tf) table 3

Table 3 Design of RSM and its actual and predict values of RSM
Run p RunOrder PtType Blocks MF(Kg/hr) MC(Kg/hr) TF LDG
(°C)
1 1 1 1 1 6.00 80.3 309.10 4.675
2 17 2 0 1 6.25 85.9 341.85 5.374
3 5 3 1 1 6.00 80.3 374.60 5.792
4 18 4 0 1 6.25 85.9 341.85 5.374
5 4 5 1 1 6.50 91.5 309.10 4.917
6 3 6 1 1 6.00 91.5 309.10 5.327
7 14 7 -1 1 6.25 85.9 374.60 5.948
8 7 8 1 1 6.00 91.5 374.60 6.600
9 12 9 -1 1 6.25 91.5 341.85 5.725
10 9 10 -1 1 6.00 85.9 341.85 5.598
11 10 11 -1 1 6.50 85.9 341.85 5.168
12 13 12 -1 1 6.25 85.9 309.10 4.801
13 6 13 1 1 6.50 80.3 374.60 5.346
14 16 14 0 1 6.25 85.9 341.85 5.374
15 20 15 0 1 6.25 85.9 341.85 5.374
16 19 16 0 1 6.25 85.9 341.85 5.374
17 11 17 -1 1 6.25 80.3 341.85 5.024
18 15 18 0 1 6.25 85.9 341.85 5.374
19 2 19 1 1 6.50 80.3 309.10 4.315
20 8 20 1 1 6.50 91.5 374.60 6.092
The best response range can be obtained by
Based on the runs given with a specific analyzing the response surface and contour
parametric values, parameters were plots.
reentered into the software (NORSOK M-
506) to calculate the response (CPR).

3. Results and Discussion

3.1Estimation of The Effect of The Three
Parameters
RSM is used to find the optimal set of the
operating parameters that produce a
minimum value of the response (LDG).
Fig. 1 Surface plot of interaction between Tf and
Mc on LDG

Figure 1 shows the three-dimensional Fig.3 Surface plot of interaction between Tf and
Mf on LDG
response surfaces for the interaction effect
of Mc and Tf on LDG at constant Mf Figure 3 shows the three-dimensional
(6.25°C). Decreasing the Mc from (92) to response surfaces for the interaction effect
(80) and decreasing Tf from 360 to 300 of Tf and Mf on LDG at constant Mc(85.9
leads to a corresponding decrease in LDG. kg/hr.) Increasing the mf from 6.00 to
6.45kg/hr. and decreasing Tf from (360o C)
to (300) leads to a corresponding decrease in
LDG.

Fig. 2 Surface plot of interaction between Mf

and Mc on LDG
Fig. 4 Contour plot of interaction between Tf
Figure 2 shows the three-dimensional and Mc on LDG
response surfaces for the interaction effect
of Mf and Mc on LDG at constant Tf Figure 4 shows a contour plot based on
(341.8o C). Increasing the Mf from 6.00 to interaction analysis between tf and Mc. The
6.45 kg/hr. and increasing Mc from 92 to 80 Mf for this analysis was set constant 6.25.
kg/hr. leads to a corresponding decrease in From this plot, we can observe that the best
LDG. value for LDG can be obtained at low Mc
value and low Tf value.
 Fig. 5 Contour plot of interaction between Mc Fig. 7 Optimizing operating parameters for LDG
and Mf on LDG through desirability function approach

Figure 5 shows a contour plot based on the 5. Conclusion

interaction analysis between Mc and Mf. The optimized values of the various input
The Tf for this analysis was set constant parameters can be summarized as follows:
(341.9o C). From this plot, we can observe Optimum Tf 3.09°C, Optimum Mc 80.3
that the best value for LDG can be obtained kg/hr and Optimum Mf 6.5 kg/hr. The
at high Mf value and low Mc value. corresponding minimum LDG obtained is
4.317.
References
1. Vinay Sati1, Anirudh Gupta Efficiency
Improvement Opportunity in Boiler Without
Changing GCV of the Coal, International
Journal of Science and Research (IJSR),
(2013), ISSN (Online): 2319-7064.
2. Moni Kuntal Bora1, S. Nakkeeran,
Performance Analysis From The Efficiency
Estimation of Coal Fired Boiler
,International Journal of Advanced Research
Fig. 6 Contour plot of interaction between Tf (2014), Volume 2, Issue 5, 561-574.
and Mf on LDG 3. R. Saidur n, J.U.Ahamed,H.H.Masjuki,
Energy, exergy and economic analysis of
Figure 6 shows a contour plot based on the industrial boilers, (2003) Elsevier Ltd.
interaction analysis between Mf and Tf. The doi:10.1016/j.enpol.2009.11.087.
Mc for this analysis was set constant (85.9). 4. Shyan-Shu Shieh a, Yi-Hsin Chang b,
From this plot, we can observe that the best Shi-Shang Jang b,*, Ming-Da Mac, Ta-Sung
value for LDG can be obtained at low Tf Huang, Statistical key variable analysis and
value and high Mf value. model-based control for the improvement
of thermal efficiency of a multi-fuel boiler,
4. Response Optimization 2009 Published by Elsevier Ltd.
The values in red represent the operating doi:10.1016/j.fuel.2009.07.001.
parameters to obtain a minimum LDG are 5. Harish Ghritlahre, Tej Pratap Singh,
shown in figure 7, and it also shows Effect of Excess Air on 30 TPH AFBC
individual parameter in each column the Boiler on Dry Flue Gas Losses and its
effect of the response and the other Efficiency, International Journal of Research
parameter kept constant. At the upper left in Advent Technology, Vol.2, No.6, June
corner, D represents composite desirability 2014 E-ISSN: 2321-9637.
while d is the individual desirability.
 6. Jian-Guo Wanga, Shyan-Shu Shiehb, Shi- Journal of Engineering and Technology. Vol
Shang Jangc, David Shan-Hill Wongc, 1. Issue 1. Jan-June 2011.
Chan-Wei Wud, A two-tier approach to the 9. A Wienese, Boilers, Boiler fuel and boiler
data-driven modeling on thermal efficiency efficiency, Proc S Afr Sug Technol Ass
of a BFG/coal co-firing boiler, 2013 Elsevier (2001) 75.
Ltd, Volume 111, doi.org/10.1016/j. fuel. 10. Einstein, Dan, Worrell, Ernst, Khrushch,
2013.03.081 Marta, Steam systems in industry: Energy
7. Jayamaha, Lal, 2008. In: Energy Efficient use and energy efficiency improvement
Building Systems, Hardbook. Mcgraw Hill potentials, 2001, Lawrence Berkeley
education, Europ. National Laboratory
8. Rahul Dev Gupta, Sudhir Ghai, Ajai Jain. https://escholarship.org/uc/item/ 3m1781f1
“Energy Efficiency Improvement Strategies
for Industrial Boilers: A Case Study”.

View publication stats