U.P.B. Sci. Bull., Series D, Vol. 73, Iss.

4, 2011 ISSN 1454-2358

A STUDY OF INJECTION TIMING FOR A DIESEL ENGINE

OPERATING WITH GASOIL AND HRG GAS

Adrian BIRTAS1, Radu CHIRIAC2

Această lucrare are ca scop evaluarea tendinţelor de modificare a

performanţelor şi emisiilor unui motor diesel cu variaţia avansului la injecţie când
motorul este alimentat în mod dual, folosind programul de simulare AVL Boost. În
acest sens modelul creat simulează funcţionarea unui motor diesel prevăzut cu un
sistem suplimentar de alimentare. Sistemul normal de injecţie pentru alimentarea cu
motorină este cuplat cu un sistem de injecţie extern care asigură îmbogătirea cu gaz
HRG (Gaz Bogat în Hidrogen) a aerului admis. Acest gaz este obţinut prin
electroliza apei şi este furnizat motorului, prin introducere în sistemul de admisie, la
diferite debite corespunzătoare procentelor energetice de H2 utilizate în amestecul
combustibil (0%, 1,46%, 3,38% si 5,85%).

This work addresses the possibility to asses the changing of performance and
exhaust emissions with the injection timing variation when a diesel engine operates
in dual fuel mode using the simulation program AVL Boost. In this sense the created
model simulates the operation of a diesel engine which is provided with a
supplementary fuelling system. The normal diesel fuel injection system is associated
with an external injection system which ensures the enrichment of the input air with
HRG - Hydrogen Rich Gas. This gas is obtained by water electrolysis and is
supplied at different flow rates to the engine intake manifold corresponding to
different energetic H2 fractions (0%, 1.46%, 3.38% and 5.85%) for the overall
mixed fuel composition.

Key words: diesel engine, dual fueling, hydrogen rich gas, injection timing,
performance, emissions

1. Introduction

Assessments on the new engines performance and emissions obtainable

are actually performed in the research and development stages using dedicated
simulation tools which offer the possibility to envisage what are the best paths to
be followed. Using computational codes for fundamental processes simulation, in
conjunction with experiments at the test bench becomes nowadays a common
1
PhD Stud., Department of Internal Combustion Engines, Faculty of Mechanical and
Mechatronics Engineering, POLITEHNICA University of Bucharest, Romania, e-mail:
[email protected]
2
Prof., Department of Internal Combustion Engines, Faculty of Mechanical and Mechatronics
Engineering, POLITEHNICA University of Bucharest, Romania, e-mail:
[email protected]
66 Adrian Birtas, Radu Chiriac

practice for the engines development activities, to obtain fast and reliable results
with low costs. The existing codes simulate the processes within cylinder (gas
flow, air-fuel mixing, combustion and heat transfer) intake and exhaust systems
(gas flow phenomena) fuel-injection systems (fuel flow and injection phenomena)
etc., in specific conditions imposed by the details of the engine component design
and operation conditions [1].
Large differences in the complexity of the mathematical models have
resulted in the existing variety of simulation codes. According to the degree of
complexity, they are based on a variable amount of input data, defined the engine
designed details, physical properties and particular coefficients (as discharge
coefficients for flows, heat transfer, fuel-air mixing and combustion). The
efficient use of the different computational codes, for reliable analysis, parametric
studies and prediction of engine performance needs a preliminary calibration
stage, based on experiments carried out at the test bed.
The AVL Boost program, v2009.1 is based on the calculation models
associated to each engine component, which are assembled in the data pre-
processor package.
According to a graphical programming method and with an interface
containing pre-defined constituent elements of the engine, one can specify the
relevant geometrical and technical features of every main part of the engine. A
symbolic model of the engine is thus designed, which is physically similar to the
investigated engine.

2. AVL Boost Model

A symbolic model of a tractor diesel engine used for experimental

research on a test bed was thus created.
The main elements of the engine symbolic model are pipes (1, 2, 3, 4, 5, 6,
7, 8, 9, 10, 11, 12 si 13) and manifolds (PL1, PL2, PL3 si PL4), junctions, air
filter (CL1), cylinders (C1, C2, C3 si C4), injector, measuring points (MP1, MP2,
MP3 si MP4) and system boundaries (SB1 si SB2). All these components need
design and operational data corresponding to the operation condition which is
investigated (Fig. 1) [2].
The fluid flows through pipes and manifolds are simulated by one-
dimensional model, with flow coefficients which have to be defined as initial data.
The flow is defined by continuity, momentum, and energy conservation equations;
the friction coefficient and the heat transfer to the walls are variable along the
pipes [3].
The cylinder is obviously the most complex part of the engine model and
it contains the most complicated associated submodels, implying thus
significantly more input data than all the other components. Connections of the
 A study of injection timing for a diesel engine operating with gasoil and HRG gas 67

cylinders with the intake and exhaust manifolds are symbolized by pipes and they
represent the ports in the cylinder head. In the case of the multicylinder engines
the manifolds are considered plenums and sometimes the connection between
ports and manifolds are symbolized by junctions. [4]
In the AVL BOOST code the engine is considered as a block of identical
cylinders, exchanging successively mass and energy with the surroundings
environment trough the valves, according to the given succession of the cylinders
operation. The engine geometric dimensions, the piston movement, the firing
order, the fuel amount, the combustion characteristic and some corresponding
estimates for the thermodynamic parameters of the cylinder charge when the
exhaust valve opening and for the heat exchange inside the cylinder must be
given as input data.

Fig. 1. The engine symbolic model created for the simulation of the HRG gas injection in the
intake manifold

Combustion process modeling for diesel engines can be obtained with a

real rate of heat release, with a formal type characteristic (single or double Vibe)
or with other type of sub-models such as: Woschni/Anisits, Hiroyasu, AVL MCC
(Fig. 2). [3]
The AVL MCC model was chosen in this study (Mixing Controlled
Combustion) for the prediction of the combustion characteristics. This model
considers the effects of the premixed (PMC) and diffusion (MCC) controlled
combustion processes according to equation (1) [5]:

dQtotal dQMCC dQPMC

= + (1)
dα dα dα
68 Adrian Birtas, Radu Chiriac

where: Qtotal total heat release [kJ];

QMCC cumulative heat release for the mixture controlled combustion [kJ];
QPMC cumulative heat release for the premixed mixture combustion [kJ] .

Fig. 2. Combustion models characterized by the type of heat release

The using of the AVL MCC combustion model, requires input data which
characterize the injection process (number and injector holes diameter, injection
pressure in fuel line, flow coefficients) and the combustion process (autoignition
delay factor, combustion parameters, for the in cylinder turbulent flow, for the
dissipated kinetic energy, for the EGR influence and for the premixed
combustion) (Fig. 3).
These sophisticated submodels are more related to the injection system
characteristics and also offer the opportunity to evaluate the formation of NOx
emissions, CO and soot. Nitric oxide emission resides in two parameters: NOx
Kinetic Multiplier-parameter to adjust the NOx producing kinetics and NOx
Postprocessing Multiplier- parameter for setting up the results of NOx generating
calculation. The carbon monoxide emission is assesed by CO Kinetic Multiplier-
parameter to adjust to CO producing kinetic. The fraction of the smoke particles is
ended by two constants Soot Production Constant - to characterize the speed of
soot formation and Soot Constant Consumption - to characterize the oxidation rate
of soot particles.
The heat transfer to the combustion chamber walls can be evaluated using
various computational submodels (e.g. Woschni 1978, Woschni 1990, Hohenberg,
Lorenz, Model AVL 2000). In this simulation study Woschni’s 1990 correlation
was selected with the following calculation relations (equation 2, 3) [3]:
Qwi = Ai ⋅ α w ⋅ (Tc − Twi ) (2)
 A study of injection timing for a diesel engine operating with gasoil and HRG gas 69

where: Qwi is the convective heat flux to combustion chamber surroundings

(cylinder head, piston, liner); Ai the corresponding surface areas, αw convective
heat transfer coefficient; Tc average cylinder charge temperature; Twi wall
temperature (cylinder head, piston, liner). [5]

Fig. 3. Input data for AVL MCC combustion model

This correlation aims to ensuring at a more accurate prediction of the heat

transfer by connecting it with the engine operation condition including therefore
the indicated mean effective pressure IMEP (equation 3):
0.8
⎧⎪ ⎡ ⎛V ⎞
2
⎤ ⎫⎪
α w = 130 ⋅ D − 0. 2
⋅p 0. 8
c ⋅T
c
− 0.53
⋅ ⎨c1 ⋅ c m ⋅ ⎢1 + 2⎜ TDC ⎟ ⋅ IMEP −0.2 ⎥ ⎬ (3)
⎪⎩ ⎢⎣ ⎝ V ⎠ ⎥⎦ ⎪⎭

where: D is the cylinder bore; pc actual cylinder pressure; c1 = 2.28 + 0.308 *

cu/cm; cm mean piston speed; cu circumferential velocity; VTDC cylinder volume
at TDC, V actual cylinder volume [5].
The normal diesel fuel injection system is associated with an external
injection system which ensures the enrichment of the input air with HRG-
Hydrogen Rich Gas. This gas is obtained by water electrolysis and is supplied at
different flow rates to the engine intake manifold corresponding to different
energetic H2 fractions (0%, 1.46%, 3.38% and 5.85%) for the overall mixed fuel
composition.
70 Adrian Birtas, Radu Chiriac

In the last years, industrial hydrogen was produced not only by methane
steam reforming but also by water electrolysis technologies which have been
continuously improved in order to reduce the energy consumption.
New type of electrolyzers based on particular technologies which do not
separate the resulting H2 and O2, gas products ( known as hydroxi gas, HHO,
Brown gas, Rhode gas) produce hydrogen-oxygen mixtures that contain hydrogen
64...67% (vol), oxygen 31...33% and 0...5% other compounds of hydrogen with
oxygen. In our case this gas named HRG gas, was considered as a stoichiometric
hydrogen-oxygen mixture. Therefore for the gas fuel supply on the symbolic
model of the engine, HRG input was delivered by the gas injector (I1 in figure 1)
and it was considered that the mass fractions of the two gases are 88.88% for
oxygen and 11.12% for hydrogen. The corresponding gas flow was continuously
injected in the intake manifold and adjusted in constant flow rates by an electronic
flow meter Alicat Scientific provided with pressure regulator [2].
Concerning the combustion process, the program needed also as input data
the mass percent of diesel fuel and hydrogen from the total amount which
participate to the combustion. These fractions are changed as function of the HRG
gas flow rate introduced in the intake manifold and as function of the diesel fuel
consumption settled experimentally on the test bed. The diesel fuel-HRG mixture
lower heating value was automatically calculated based on these fractions and on
the lower heating values of the each fuel (diesel and hydrogen), taken from the
AVL Boost program data base.
The simulations were based on the hypothesis of internal mixture
formation due to the in-cylinder direct injection of the diesel fuel. Thus it was
assumed that:

- fuel added to the cylinder charge is immediately combusted;

- combustion products mix instantaneously with the rest of the cylinder
charge and form a uniform and homogeneous mixture;
- as a consequence, the cylinder charge A/F ratio diminishes continuously
from an initial value at the start of combustion to a final value at the end of
combustion.

3. Engine test bed

The experimental investigation, for this simulation study was performed at

the University Politehnica of Bucharest on a tractor diesel engine, made by Uzina
Tractorul Brasov. The engine was internationally certified in Czech Republic in
2004. The engine technical specifications are listed below:
 A study of injection timing for a diesel engine operating with gasoil and HRG gas 71

Engine model: 2404.050

Number of cylinders: 4 line
Engine capacity: 3758 cm3
Number of valves: 8
Power: 50 kW at 2400 rpm
Naturally aspirated
Maximum torque: 228 Nm at 1100 rpm
Direct Injection
Relative air/fuel ratio: 1.4
Compression ratio: 17.5
Bore: 102 mm
Stroke: 115 mm
Connecting rod length: 182 mm

The engine was mounted on a test bench specially provided with dedicated
equipment for accurate measurement and control of the operating condition
parameters. Engine loaded was performed by an eddy current dynamometer type
AVL ALPHA 160. Fuel consumption was gravimetrically measured by an AVL
Dynamic Fuel Meter 733 S coupled with an AVL Fuel Temperature Controller
753 C. Exhaust gas emissions were determined using a gas analyzer HORIBA
MEXA 7170 D. Cylinder pressure traces have been registered by means of an
AVL Indiset 620 supported by the AVL Indicom 1.6 software.
The testing matrix was fitted-out by the automatic regulating and control
system AVL Puma Open. This test bench management system facilitated the
accomplishment of all stages of the testing matrix, with a fast response time and
allowed in the same time the primary data processing.

4. Model calibration on experimental data

Before starting intensive parametric studies to perform engine

optimization, a careful calibration procedure of the engine model is always
necessary. Almost every time, it is necessary to perform more than one simple
calibration check point in order to increase the model prediction accuracy.
The model calibration is normally based on the cylinder pressure trace and
on as much possible measured parameters defining the engine operation regime.
These parameters are experimentally obtained by test bed investigation and they
may be: power, torque, air and fuel consumptions, volumetric efficiency, air-fuel
ratio, residual burned gases fraction, local pressures and temperatures measured in
different points of interest along the intake and the exhaust system. System
boundary conditions are also extremely important because they imposed the
72 Adrian Birtas, Radu Chiriac

relation between the engine performance and emissions parameters and the
ambient [6].
Therefore, local temperatures and pressures measured on the testing bed
are compared with the AVL Boost program results for different locations:

- intake manifold - Measuring Point 1 – MP1(in Boost) and T_admis

and P_admis (on the testing matrix)
- exhaust manifold - Measuring Point 3 – MP3 (in Boost) and T_EXH_1
and P_EXH_1 (on the testing matrix)
- emissions gas analysis sample probe position– Measuring Point 4 –
MP4 (in Boost) and T_EXH and P_EXH (on the testing matrix).

Fig. 4 and table 1 show that pressures and temperatures calculated with
AVL Boost, and presented with AVL IMPRESS Chart are comparable with the
measured and registered averages values by AVL Puma Open on test bed. These
data figure the measured and calculated values for the engine operation condition
at part load 60%, speed 2400 rpm, and for 5.85% H2 energetic fraction of HRG in
fuel mixture [2].

Fig. 4. Local temperatures and pressures calculated with AVL Boost in different locations for
engine operating condition 60% load, 2400 rpm, and 5.85% H2

Comparison of the experimental pressure trace (as an average diagram for

300 consecutive cycles registered) with the resulted curve obtained from AVL
 A study of injection timing for a diesel engine operating with gasoil and HRG gas 73

Boost simulation is shown in Fig. 5. This comparison shows a good agreement

between experiment and simulation and an accurate model calibration.
Table 1
Local temperatures and pressures measured on test bed in the same locations and for the
same engine operating condition
P_admis P_EXH P_EXH_1 T_admis T_EXH T_EXH_1
Pa Pa Pa K K K
-103000 100250 122000 298 564 651

60

50

Tes ting bed

40 A V L  B oos t
bar

30

20

10

0
‐100 ‐80 ‐60 ‐40 ‐20 0 20 40 60 80 100

°C A

Fig. 5. Pressure traces comparison for model calibration at 60% load, 2400 rpm, and 0% H2

The study of the effect of the injection timing is accomplished for the run
at 2400 rpm and 60% load. For each of the energetic additive proportions with
HRG pursued ( 0%, 1.46%, 3.38% and 5.85% ) calibration of the model pursuing
the experimental data was undertaken.
It can be noticed that the relative deviations of the simulation results face to
the reference experimental data are less than 0.25% for effective power, 2.9% for
maximum pressure and 5.15% for brake thermal efficiency BTE (table 2 and 3).
After the calibration procedure is finalized, the model can be used for
engine the optimization activity based on parametric studies. Many design and
operational parameters can be modified, as for example valves timing and valves
lift curves, combustion chamber geometry, injection timing and injection system
characteristics, air-fuel ratio, heat release characteristics etc.
In this study, the influence of diesel fuel injection timing variation on the
engine performance and pollutant emissions formation was considered. The
injection characteristic was modified by changing the injection timing, the total
injection duration and the normalized injection rate being unchanged as it can be
seen in Fig. 6.
74 Adrian Birtas, Radu Chiriac

Table 2
Model calibration results for 0% and 1.46% HRG
Measured Simulated Relative deviation %
H2 energetic fraction [%] 0% 1.46% 0% 1.46% 0% 1.46%
Effective power [kW] 26.84 26.84 26.84 26.79 0.00 0.19
Diesel fuel consumption [kg/h] 7.66 7.54 7.66 7.54 0.00 0.00
Lambda 2.39 2.43 2.39 2.43 0.00 0.00
Maximum cylinder pressure [bar] 53.05 53.80 52.64 52.25 0.77 2.88
NOx emission [ppm] 345 370 345 370 0.00 0.00
CO emission[ppm] 336 302 336 302 0.00 0.00
Specific smoke value [g/kWh] 0.1957 0.1870 0.1953 0.1860 0.21 0.52
BTE [%] 29.75 29.66 28.22 28.19 5.12 4.96
Indicated efficiency [%] 38.02 37.91 38.50 38.49 -1.26 -1.52

Table 3
Model calibration results for 3.38% and 5.85% HRG
Measured Simulated Relative deviation %
H2 energetic fraction [%] 3.38% 5.85% 3.38% 5.85% 3% 5.85%
Effective power [kW] 26.84 26.84 26.78 26.78 0.22 0.22
Diesel fuel consumption [kg/h] 7.41 7.23 7.41 7.23 0.00 0.00
Lambda 2.45 2.49 2.45 2.49 0.00 0.00
Maximum cylinder pressure [bar] 53.84 54.06 52.31 53.28 2.84 1.44
NOx emission [ppm] 386 406 386 406 0.00 0.00
CO emission[ppm] 288 264 288 264 0.00 0.00
Specific smoke value [g/kWh] 0.1505 0.1423 0.1507 0.1424 -0.15 -0.05
BTE [%] 29.59 29.57 28.12 28.07 4.99 5.05
Indicated efficiency [%] 38.08 37.88 38.38 38.32 -0.78 -1.16

Fig. 6. Normalized injection rate as an unchanged input parameter

 A study of injection timing for a diesel engine operating with gasoil and HRG gas 75

5. Injection timing effects

The AVL Boost program adapted for engine operating conditions with air
enriched HRG in the intake system was used to study the effects of diesel fuel
injection variation on performance and emissions. The study tried to evaluate the
possibility to compensate the decrease in engine efficiency determined due to
HRG addition, which was experimentally established by means of injection
timing optimization.
The parametric study of the influence of injection timing variation was
accomplished for different values ranging from 4°CA to 22°CA. The main results
of the simulation are figured below in Figs. 7, 8 and 9.

40

39

Indicated
38
Power
kW

0% HRG
37
1.46% HRG
3.38% HRG
36 5.85% HRG

31

30
Effective
29 Efficiency
%

0% HRG
28 1.46% HRG
3.38% HRG

27 5.85% HRG

30

29.5

29

28.5
Effective
kW

28 Power
27.5 0% HRG
27 1.46% HRG
3.38% HRG
26.5
5.85% HRG
26

25.5
4 6 8 10 12 14 16 18 20 22
Injection Timing [ºCA]

Fig. 7. The influence of injection timing on indicated power, effective efficiency and effective
power at 60% load and 2400 rpm
76 Adrian Birtas, Radu Chiriac

The following aspects can be noticed:

The increase of the injection timing from 6 °CA to 8 °CA, for a constant
HRG flow rate corresponding to 3.38% H2, leads to an increase of the effective
power by 3.8%, of the maximum pressure by 11.3 % and of the brake thermal
efficiency by 1.92 %; these effects can be attributed to the combustion
improvement due to its shifting towards top dead center.
The operation with 3.38% H2 and injection timing 7 °CA relative to the
reference condition (0% H2, and injection timing 6 °CA), is associated with a
decrease of the smoke number by 25%, of the CO emission by 17%, and with an
increase of the NOx emission by 24%. The maximum cylinder pressure records a
moderate increase from 52.64 bar to 55.27 bar.

90

80
pm ax
70
0% HRG
1.46% HRG
bar

60
3.38% HRG
5.85% HRG
50

3.5

3 (dp/dα)m ax
0% HRG
2.5
1.46% HRG
bar/deg

3.38% HRG
2
5.85% HRG
1.5
140

138

136

134
Δαa
deg

132 0% HRG
1.46% HRG
130
3.38% HRG
128 5.85% HRG

126

124

122
4 6 8 10 12 14 16 18 20 22
Injection Timing [ºCA]

Fig. 8. The influence of injection timing on maximum pressure, maximum pressure rise and on
total combustion duration at 60% load and 2400rpm
 A study of injection timing for a diesel engine operating with gasoil and HRG gas 77

500
450
400
NOx
350 0% HRG
300 1.46% HRG
ppm

3.38% HRG
250
5.85% HRG
200
150
380

340 CO
300 0% HRG
1.46% HRG
ppm

260
3.38% HRG
220 5.85% HRG

180

140
0.21
Soot
0% HRG
0.17
g / kWh

1.46% HRG
3.38% HRG
0.13 5.85% HRG

0.09

0.05
4 6 8 10 12 14 16 18 20 22
Injection Timing [ºCA]

Fig. 9. The influence of injection timing on NOx, CO and smoke, at 60% load and 2400rpm

6. Conclusions

In conclusion, the simulations performed with the AVL Boost program

have showed that for reduced HRG flow rates and optimized diesel fuel injection
timing, itcan be obtained:

1. Slight increase in the engine brake thermal efficiency and moderate increase of
the maximum cylinder pressure.
2. Significant smoke and CO emission reductions, but important NOx emission
increase, which however can be reduced by promoting the cooled EGR solution.
3. These aspects seem to be particularly interesting from the perspective of
modifying the injection characteristic of the Delphi rotary type injection pump, in
direction of advancing the static injection timing.

Acknowledgments:
The authors acknowledge the AVL – Advanced Simulation Technologies
team for its support in performing this study.
78 Adrian Birtas, Radu Chiriac

REFERENCES

[1] R. Chiriac, G. Descombes, Fuel consumption and pollutant emissions reduction for diesel
engines by recovery of wasted energy, Enviromental Engineering and Management Journal,
vol 9, nr.10, 2010, “Gheorghe Asachi ” Technical University Iasi.
[2] A. Birtas An investigation of the possibility to improve with HRG gas emissions and efficiency
of a diesel engine, PhD Thesis, Politehnica University of Bucharest, Bucharest, feb 2011.
[3]*** AVL List Gmbh, AVL Boost – User`s Guide, 2000.
[4] A. Birtas, C. Petcu, C. Dica, R. Chiriac, On the possibility to predict the performance of a
tractor diesel engine using AVL Boost simulation code, The 33rd International Conference
on Automotive Engineering , Off Road Vehicles - ORV2008503, Bucharest 25-26 sept
2008.
[5]*** AVL List Gmbh, AVL Boost – Theory, 2009.
[6] R. Chiriac, Diagrama indicată – Mijloc clasic şi modern de optimizare a funcţionării
motoarelor cu ardere internă, (Indicated diagram - Classic and modern way of optimizing
the operation of internal combustion engines) Revista TERMOTEHNICA, nr. 1-2/2003,
pag. 55-61, ISSN 1222-4057, 2004. (in Romanian)