experimental investigation on attenuation of emission with optimized lpg jet induction in a dual fuel diesel engine and prediction by ann model

The optimized jet parameters obtained through CFD technique and the experimental results achieved from the dual fuelled CI engine show considerable improvement in the engine performance

Energy Procedia 14 (2012) 1427 – 1438

1876-6102 © 2011 Published by Elsevier Ltd Selection and/or peer-review under responsibility of the organizing committee of 2nd International Conference on Advances in Energy Engineering (ICAEE).

doi:10.1016/j.egypro.2011.12.887

Available online at www.sciencedirect.com

Energy Procedia

Energy Procedia 00 (2011) 000–000

www.elsevier.com/locate/procedia

Available online at www.sciencedirect.com

Advances in Energy Engineering

Experimental Investigation on Attenuation of Emission with

Optimized LPG Jet Induction in a Dual Fuel Diesel Engine

and Prediction by ANN Model

Thomas Renald C.Ja*, Somasundaram Pb

a Department of Aeronautical Engineering, Sri Ramakrishna Engineering College, Coimbatore, Tamilnadu, India

b Department of Mechanical Engineering, KS Rangasamy College of Technology, Tiruchengode, Tamilnadu, India

Abstract

Environment pollution due to vehicles exhaust emission is still a severe crisis and an international concern has been raised for its control and diminution Especially, man-kind is experiencing an epidemic of illnesses made worse by air pollution mainly because of increased number of automotive vehicles The main problem faced by most of the vehicles today is the emission of NO x which can be controlled by different ways such as exhaust gas recirculation, using alternate fuels, turbo charging, and different mode of combustion On the other hand, diesel engines are the most trustworthy power sources in the transportation Due to inflexible emission norms and rapid depletion of petroleum resources, there has been a continuous exertion to use alternative fuels Further researches are being carried out to reduce emission rates and these researches would be explored till reducing the exhaust emission upto zero level In such a way, the present investigation explores with a series of experiments towards effective combustion of air, LPG and diesel mixture without encompassing major modification in the engine construction Here, LPG jet is inducted through air inlet to accomplish homogeneous mixture then it is allowed into the combustion chamber of CI engine The major parameters of the LPG jet injector are optimised through CFD technique by utilizing a commercial CFD code Changes in the performance of the engine and emission levels with the influences

of the jet parameters are observed and analysed experimentally The optimized jet parameters obtained through CFD technique and the experimental results achieved from the dual fuelled CI engine show considerable improvement in the engine performance and significant reduction in CO 2 , CO, HC and NO x emission besides the normal engine

performance and emission levels of diesel fuel induction Eventually, an artificial neural network (ANN) model is

developed for predicting the emission levels based on the jet parameters, applying load and % of LPG induction by utilizing the experimental results It is also found that the predicted results provided good agreement with the experimental results

© 2011 Published by Elsevier Ltd

Keywords: Dual fuel CI engine, LPG jet injection, homogeneous air-fuel mixture, jet parameters, CFD approach, reduction of

emission, engine performance, ANN model

* Corresponding author Tel.: +919894956436 E-mail address: thomsi_reni2000@yahoo.co.in

© 2011 Published by Elsevier Ltd Selection and/or peer-review under responsibility of the organizing committee

of 2nd International Conference on Advances in Energy Engineering (ICAEE)

1 Introduction

LPG known as an auto gas is a mixture of Propane and Butane LPG is probably the third largest used fuel after petrol and diesel and is widely used as a cleaner, eco-friendly automotive fuel

LPG produces significantly less carbon monoxide and oxides of nitrogen emissions as well as a smaller percentage of carbon dioxide emissions LPG also emits 90% less particulates than diesel engines Reductions in emissions from LPG have resulted in the government, offering a variety of incentives to encourage motorists to convert to LPG Car owners driving vehicles using alternative fuels will be taxed less than vehicles using petrol or diesel Other environmental advantages of Liquid Petroleum Gas as an alternative fuel include the fact that LPG engines are significantly quieter than diesel engines and marginally quieter than petrol engines Hence there is a scope for designing a gas jet and its position to study the engine performance rate and emission Gisoo Hyun et al [1] developed an LPG fuelled direct injection SI engine, especially in order to improve the exhaust emission quality while maintaining high thermal efficiency comparable to a conventional engine through Computational Fluid Dynamics (CFD) Chunhua Zhang et al [2] conducted a study on the control scheme of a liquefied petroleum gas (LPG)– diesel dual-fuel engine with electronic control The experimental results showed that comparing with diesel, the output performance of dual fuel is not reduced, while smoke emission of dual fuel is significantly reduced, NOx emission of dual fuel is hardly changed, but HC emission and CO emission of dual fuel are increased and fuel consumption of dual fuel is reduced Kurniawan et al [3] used computational fluid dynamics (CFD) simulation and investigated the effect of piston crown shape to air motion characteristics of an internal combustion engine White et al [4] used CFD packages Fluent to model the direct-injection of two such fuels simultaneously into an engine and studied about the salient features of the two fuel jets are being to optimize the design of a dual-fuel injector for compression-ignition engines Roberto et al [5] examined the flow geometry effects on the turbulent mixing efficiency quantified as the mixture fraction and compared two different flow geometries are at similar Reynolds numbers, Schmidt numbers, and growth rates, with fully developed turbulence conditions Ren et al [6] investigated the performances of the gaseous fuel supply and its influence on hydrocarbon (HC) emissions of dual-fuel engines and presented a new design of manifold respirators with mixers Cao et al [7] analysed an experiment on engine performance and sprays and characteristics between diesel and mixed liquefied petroleum gas (LPG)/diesel injection engines The performance test results showed that with LPG the mixed ratio increases, engine power reduces slightly, fuel consumption and engine noise have almost no change, pollutant emissions of smoke, CO and NOx at full load are improved significantly, but the amount of unburned HC increases

In the above literature survey it has been reviewed that the mixture formation is good by modifying the design of piston head The performance of the engine increases when the LPG is electronically injected with micro controllers according to the speed and power required The emissions CO and NOx from the dual fuel engine are greatly reduced HC emissions from dual fuel engine are par with that of diesel at full load and its less in low loads and half loads The performance of the engine could be increased by designing the piston head by taking swirl and tumble ratio of the entering air Studying the flow characteristics and changing the flow pattern shows a significant change in performance of engine The design of the gaseous fuel supply system has a great influence on HC emissions in dual-fuel engines

at light load A manifold respirator with a mixer gives the best performance in reducing HC emissions compared with a common pipe mixer and a respirator with no mixer Literature survey confirms that CFD tool can be used to study the gas jet parameters so as to improve performance of the engine and emission Major problems in emission control are design of engine and improper air-fuel mixture in the combustion system In an ideal combustion process, air-fuel mixture would be burnt completely within the combustion chamber In actual, when fuel is burnt, it emits virtually insignificant CO, O2 and relatively little NOx, the main constituents of acid rain, and substantially less CO2, a key culprit in the greenhouse debate, than most oil products and coal

Thomas Renald C.J and Somasundaram P\ / Energy Procedia 14 (2012) 1427 – 1438 Author name / Procedia Environmental Sciences 00 (2011) 000–000 1429 Major techniques can be followed in the reduction of emission problem are,

• Modifying the engine design;

• Using alternate fuel which produces low emission;

• Recirculating or treating the exhaust for further utilization or conversion

The emission control problems have been addressed in numerous research works, because reduction

of emission is a difficult problem which is influenced by design of an engine, appropriate fuel, proper

air-fuel ratio and homogeneous mixture, the engine operating parameters include load, speed,

compression ratio, pilot fuel injection timing, pilot fuel mass inducted and intake manifold conditions

Finally, it has been found that there is a scope for improving the performance and reducing emission level

of diesel engines exclusive of major modification in the engine construction

Nomenclature

d inlet diameter of nozzle in mm

x distance between tip of the nozzle and intersection of vertical axis of air inlet pipe and

nozzle axis in mm

θ inclination of nozzle in degree

CO2 Carbon Dioxide

CO Carbon Monoxide

FC Fuel Consumption

FHP Frictional horse Power

IP Indicated Power in W

NOx Nitrogen Oxide

RPM Revolution per Minute

SFC Specific Fuel Consumption

ηmech Mechanical Efficiency

ηB.TH Brake Thermal efficiency

ηI.TH Indicated Thermal Efficiency

2 Problem statement

Recent researches and investigations on the compression ignition (CI) engine performance parameters

optimization and emission problems are mainly dealing with modification of engine construction to

improve the performance and reduction of emission which is tedious and non-interchangeable The

present study made an attempt on reduction of emission level in the CI engine without having any major

alteration in the existing engine construction and also to improve the performance of engine In the

betterment of engine efficiency and reduction of emission level, LPG and air should be mixed

homogenously The main objective of this work is to design a simple and realistic setup which would not

disturb the existing engines construction and the developed setup could be adapted to all engines

efficiently

3 Method

Fig.1 sho

the accomp

homogeneou

achieve this

is formidabl

problem, C

conditions

which gove

utilizing a c

Various cas

triangular el

configuratio

from this inv

inclination a

ID to see th

follows

From the

LPG jet affe

conducted f

vertical axis

loading con

engine

Fig 2: a) Grid ge

b) Contour of pat

of Analysis

ows the meth plishment of

us mixture, b , LPG jet is in

le to analyse a omputational

F inite Volum

ern the turbul commercial C ses and confi lements were ons and positio vestigation is and inserted b

he extent of

1 Velocity

2 Pressur

3 The sur

4 The sur

5 The end

e results obtai ect the homog for various in

s of air inlet pi nditions to inv

eneration when n

th lines colored b

od of analysis efficient co before the mix nducted again and decide the

F luid Dynam

me Method (F

ent flow in t CFD code k-ω figurations we used to mod ons of LPG in given in Fig

by 10mm cond turbulence pr

y inlet conditi

e inlet was giv rrounding face rrounding of in

d of LPG duct

ined by simul geneous mixin nlet diameter ipe and nozzle vestigate the i

nozzle is of 5mm

by particle ID

s and as per th mbustion, ai xture enters i nst the flow of

e condition of

mics (CFD) t

FVM) was ad the presence

ω Shear Stres ere modelled del both air an nductor in the

2 Fig 2a sho dition Fig 2b roduced at th ion was given ven for LPG i

es were consid nlet duct was

t was consider

lation, it is re

ng of air/LPG

of nozzle, dis

e axis, inclinat influences of dia at 30° inclina

his method on

r and LPG into combusti

f inlet air Bef

f LPG jet flow technique wa dopted to solv

of LPG jet lo

ss Transport m

to achieve h

nd LPG flow

e cases of noz ows grid gene

b presents the his condition

n to air inlet

inlet (Operatin dered as wall

taken as wall

red outflow

ealized that th

G Based on th stance betwee tion of nozzle these parame ation and inserted

nly the presen are premixe ion chamber fore conductin

w to be induct

as approached

ve Navier-Sto ocated in the model was us homogeneous path Tests w zzle and duct

eration when n contour of pa The boundar

ng pressure =

e physical par hese results, a

en tip of the

e, % of LPG an eters over em

d by 10mm,

nt study was c

ed together t

to m with di

ng series of ex ted In order to

d to fix the L okes and ener flow path o sed as a turbu

s mixture of were conducte The better re nozzle is of 5m ath lines color

ry conditions 2.1MPa)

rameters of n

a series of exp nozzle and in

nd % of diese mission and pe

compassed In till achieving iesel fuel To xperiments, it

o analyse this LPG jet flow rgy equations

f air inlet by ulence model air/LPG and

ed for various esult obtained

mm dia at 30°

ed by particle are given as

nozzle and the periments was ntersection of

el for different erformance of

n

g

o

t

s

w

s

y

d

s

d

°

e

s

e

s

f

t

f

Thomas Renald C.J and Somasundaram P\ / Energy Procedia 14 (2012) 1427 – 1438 1431

4 Experim

The exp

introduced a

mixture of L

such as inlet

air inlet pipe

conditions o

Different no

were adopte

were taken

100%, 20%

kg, 5kg, 10k

4.1 Experim

In the pr

as 1:0, 0.8:0

conditions a

speed of the

NOx were ex

CASE 1: 10

In this ca

CASE 2: 20

Here, the

at different

results were

CASE 3: 40

With 40%

different ang

Fig 3 Sch

ental setup

perimental set

at the centre o

LPG and air b

t diameter of

e and nozzle a

over emission

ozzles with inl

ed as 5mm, 7m

as -180°, 30°

and 80%, 40

kg and 15kg T

mental Method

resent study, a

0.2, 0.6:0.4 an

as mentioned

e engine main

xamined with

00% DIESEL

ase, the engine

0% LPG AND

e engine was

‘x’ values wit

obtained

0% LPG AND

% of LPG and

gle of inclinat

hematic of exper

Author name /

tup developed

of elbow port before it enters

nozzle (d), di axis (x), inclin

and performa

let diameter (d

mm, 8mm and

°, 45° and 90

% and 60%, a The schematic

dology

a series of exp

nd 0.4:0.6 for above The ex ntained consta the help of se

e was operate

D 80% DIESEL tested with 20

th different an

D 60% DIESEL

d 60% of dies tion for variou imental setup

/ Procedia Enviro

d for the pre tion of air inle

s into engine stance betwee nation of nozz ance of engine

d) of 5mm, 7m

d 10mm, base

°; different % and 60% and

c of nozzle arr

periments was

r various ‘d’ v

xperiment wa ant at 1500 rp eparate appara

d with 100%

L 0% of LPG an ngle of inclina

L sel for various

us loading con

onmental Sciences

esent investig

et pipe to inje cylinder To

en tip of the n

le (θ), % of L

e, the experim

mm, 10mm an

ed on the simu

% of LPG and 40% respecti rangement wit

s carried out f

values, ‘x’ val

as conducted b

pm The exhau atus called exh diesel fuel and

nd 80% of die ation for vario

s inlet diamet nditions, the em

Fig 4

s 00 (2011) 000–

ation is show ect LPG jet fo investigate th nozzle and inte LPG and % of mental work w

nd 12mm wer ulation results

d % of diesel ively for vario

th air inlet pip

for different fu

lues and ‘θ’ v

based on four ust gases such haust gas anal

d the emission esel for variou ous loading co

ter of nozzle, mission result

4 Nozzle arrangem

–000

wn in Fig 3

or achieving h

he influences o ersection of v diesel for diff was designed in

re fabricated;

s, nozzle incli combination ous loading co

pe is shown in

uel ratio of di values for diff cases as liste

h as CO, CO2 lyzer ‘AVL 43

n results were

us inlet diame onditions and

at different ‘x

ts were accom ment with air inle

Nozzle was homogeneous

of parameters ertical axis of ferent loading

n such a way

the ‘x’ values

ination angles

s are 0% and onditions of 0

n Fig.4

esel and LPG ferent loading

ed below with

2, O2, HC and 37C5’

e acquired

eter of nozzle,

d the emission

x’ values with

mplished

et pipe

s

s

s

f

g

s

s

d

0

G

g

h

d

,

n

h

CASE 4: 60

In this c

nozzle, at d

emission res

5 Results a

5.1 Engine p

Figures f

consumption

loading con

60% of dies

improved br

mechanical

efficiency is

plot of indi

efficiency is

case 60% of

Fig 5 L

Fig

0% LPG AND

case, the engin

different ‘x’ va

sults were exa

and discussion

performance

from 5 to 8 sh

n, brake therm

ditions respec

el and 40% of rake thermal e

efficiency for

s clearly notic

icated therma

s observed for

f diesel and 40

Load Vs SFC

g.7 Load Vs Mec

D 40% DIESEL

ne was opera alues with dif amined

n

how the engin mal efficiency, ctively From

f LPG Fig 6 efficiency is n

r different load ced for the cas

al efficiency f

r the case 60%

0% of LPG sh

chanical efficienc

L ated 60% of L fferent angle

ne performanc , mechanical e Fig 5, reduce depicts brake noted for the ding condition

se 60% of die for different

% of diesel an hows better en

cy

LPG and 40%

of inclination

ce based on th efficiency and

ed specific fu

e thermal effic case 60% of

ns is provided esel and 40%

loading cond

nd 40% of LP ngine performa

Fig.6

Fig

% of diesel for

n for various l

he major param

d indicated the uel consumptio ciency for vari diesel and 40

d in Fig 7 and

of LPG Fig

ditions and en

PG On the w ance

Load Vs Brake t

8 Load Vs Indic

r various inle loading condi

meters such as ermal efficienc

on is observed ious loading c 0% of LPG C

d the maximum

8 presents th nhanced indic whole, it is obv

thermal efficiency

cated thermal effi

et diameter of itions and the

s specific fuel

cy for various

d for the case conditions and Comparison of

m mechanical

he comparison cated thermal vious that the

y

iciency

f

e

l

s

e

d

f

l

n

l

e

Thomas Renald C.J and Somasundaram P\ / Energy Procedia 14 (2012) 1427 – 1438 Author name / Procedia Environmental Sciences 00 (2011) 000–000 1433

5.2 Emission level

The emission results of main toxic gases such as CO, CO2, HC, NOx and amount of O2 are compared

and presented in Figures from 9 to 13 for different loading conditions of 0 kg, 5 kg, 10 kg and 15 kg Fig

9 shows the comparison plot of CO %vol for different diesel and LPG fuel ratio with different loading

conditions and minimum CO % vol is noted for the case 80% of diesel and 20% of LPG The comparison

plot of CO2 % vol in the exhaust gas for various diesel and LPG fuel ratio with different loading

conditions is presented in Fig 10 From this figure, reduced concentration of CO2 %vol is observed for

the case 40% of diesel and 60% of LPG The level of O2 % vol is compared and shown in Fig 11 and

minimum concentration of O2 % vol is noticed for the case 60% of diesel and 40% of LPG which implies

that air/LPG mixture is homogeneous and air-fuels mixture is burnt fully There is no considerable

change in the concentration of HC which can be seen from Fig 12 Fig 13 delineates the comparison of

concentration level of NOx for various diesel and LPG fuel ratio with different loading conditions and

predominant reduction in NOx emission is examined for the case 40% of diesel and 60% of LPG

According to these results, it is obvious that increase in LPG fuel ratio decreases the emission rate

considerably

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

Load in kg

0 0.5 1 1.5 2 2.5 3

Load in kg

14

15

16

17

18

19

20

Load in kg

0 100 200 300 400 500 600 700

Load in kg

100% DIESEL 80% DIESEL+20% LPG 60% DIESEL+ 40% LPG 40% DIESEL+60% LPG

From the

by 10mm d

homogeneou

conducted w

of nozzle an

performance

in emission

of the case

performance

observed th

proportions

there is a ne

with reduce

artificial neu

e simulation r

distance insid

us mixture o

with some pos

nd different f

e was observe

level of vario

es which imp

e for nozzle d

hat nozzle co

and nozzle co

ed to optimize

ed emission ra

ural network (

roposed Art

results, it is ob

de through elb

of air and LP sible cases an fuel proportio

ed for the case ous gases are plies that inc diameter of 5 onditions infl onditions are

e these param ate In such a (ANN) model

tificial Neu

Fig.14 Prop

-20 0 20 40 60 80 100 120 140

0

Fig.13 Lo bserved and d bow of air in

PG fuel Acc

nd combination ons From the

e 60% diesel a examined for crease in LPG

mm at x =10 luence the en confined with meters to fix an

a way, the cur

ural Networ

posed ANN arch

5

100% DIESEL 60% DIESEL+40%

oad vs NO x

decided that th nlet pipe at a cording to thi

ns to investiga

e experimenta and 40% LPG

40% diesel a

G decreases 0mm subtende ngine perform hin the above

n exact conditi rrent problem

rk

itecture with a sin

Load in kg

80%

he nozzle with

an angle of

is result, a s ate the influen

al results and On the contr and 60% LPG emission lev

ed angle of 30 mance and e

e discussed an ion to achieve

m was predicte

ngle hidden layer

% DIESEL+20% LPG

% DIESEL+60% LPG

h diameter of 30° provides series of expe nces of physic

d discussion, ary, considera fuel combina vel but affect 0° in both ca emission leve

nd mentioned

e better engine

ed and optimi

r

5mm inserted predominant eriments was cal parameters better engine able reduction ation for most

ts the engine ases It is also

el But fuel range Hence

e performance ized by using

d

t

s

s

e

n

t

e

o

l

e

e

g

Thomas Renald C.J and Somasundaram P\ / Energy Procedia 14 (2012) 1427 – 1438 Author name / Procedia Environmental Sciences 00 (2011) 000–000 1435

In order to predict and optimise the present problem, an ANN based model was developed with the

base of feed forward back propagation method A code was developed based on the feed forward back

propagation method by utilizing Matlab R2009a software package Fig 14 shows the developed

architecture of artificial neural network with five input layer with 5 neurons, seven output layer with 7

neuron and one hidden layer with 10 neurons

The links with synaptic weights are connected between neurons and the back-propagation training

algorithm is based on weight updates so as to minimize the sum of squared error for K-number of output

neurons, given as

2

1( . )

2

1

p K

E= ∑ −

= (1)

where d k,p = desired output for the p th pattern The weights of the links are updated as

) )

)

1

where n is the learning step, η is the learning rate and α is the momentum constant In equation (4), δ pj

is the error term, which is given as follows:

For output layer : δpk = (d kp−o kp)( 1 −o kp),k= 1 , K (3)

For hidden layer :δpj =o pj( 1 −o pj) ∑ δpk w kj,j= 1 , J (4)

where J is the number of neurons in the hidden layer The training process is initialized by assigning

small random weight values to all the links The input–output patterns are presented one by one and

updating the weights each time The mean square error (MSE) at the end of each epoch due to all patterns

is computed as

∑ ∑

p

k

k d kp o kp NP

MSE

1 1

2

) (

1 (5)

where NP =number of training patterns

The training process will be terminated when the specified goal of MSE or maximum number of

epochs is achieved The activation function for the input and the one hidden layer is chosen as

tansigmoidal function The activation function for the output layer was chosen as pure linear function

The network was then simulated for the input values and a graph is plotted between the output and target

(neural network output) values The network created was trained for the input and output values The

stopping criterion for training was number of epochs and was given as 1000 as shown in Fig.15

The network was again simulated for the input values and the target values of the experiments

conducted The input values for the test readings were then given and the network was trained The

target value was then obtained and compared with actual outputs The results were compared with the

actual experimental results and the predicted results obtained from the present study show minimal in

variations From Fig 16, it is clear that the parameters considered could be confidently used for the above

method for predicting the emission rates The behaviours of the parameters are also noted

The predicted emission levels were compared with the respective experimental results and the

absolute percentage error was computed, which is given as

alvalue Experiment

ue edictedval alvalue

(6)

Author name / Procedia Environmental Sciences 00 (2011) 000–000

The experimental results of emission levels were utilized for predicting the emission level with the

influences of ‘d’ values, ‘x’ values, ‘θ’ values for different proportions of fuels (Diesel+LPG) with

different loading conditions Both results were compared and shown through Table 1, the absolute percentage error ranges from 0.11498% to 3.28939 % which is in the acceptable range

7 Conclusions

From all the above results and discussion, the following conclusions are arrived and they are summarized as follows:

• It was found that the new technique which was used in this investigation away from the engine construction influenced on the engine performance and emission levels predominantly

• The thermal efficiency of the engine increases when powered with dual fuel The percentage increases about 5% when 60% of diesel and 40% of LPG

• No considerable change was found in torque and brake mean effective pressure

• The mechanical efficiency was increased about 5% when powered with dual fuel when 60% of diesel and 40% of LPG

• There was an increase in Hydrocarbon emission when going to dual fuel mode

• Specific fuel consumption was reduced around 33%

• NOx was reduced upto 35% for 60% LPG and 40% diesel proportion when compared to that of running at 100% diesel

• CO2 emission was reduced about 67% for 60% LPG and 40% diesel proportion

• The CO emission was reduced upto 12% in dual fuel mode for 60% of diesel and 40% of LPG

• The absolute percentage error between experimental and predicted results ranges from 0.11498%

to 3.28939 % which is in the acceptable range

• All these results were obtained for the effective nozzle condition, diameter of 5 mm, x= 10 mm and θ = 30°

• On the whole, it is recommended that better engine performance and reduced emission level can

be achieved if the dual fuels proportion lies between 40-60 % of diesel and LPG for the nozzle condition, diameter of 5 mm, x= 10 mm and θ = 30°