Communications and Network, 2013, 5, 53-59
doi:10.4236/cn.2013.52B010 Published Online May 2013 (http://www.scirp.org/journal/cn)
Robustness and Accuracy Test of Particular Matter
Prediction Based on Neural Networks
Jiamei Deng1, Shaohua Zhong2, Andrew Ordys1
1School of Mechanical and Automotive Eng., Kingston University London, UK SW15 3DW
2School of Automotive Engineering, Wuhan University of Technology, Wuhan 430070, China
Email: firstname.lastname@example.org, email@example.com,
The increasingly stringent emissions regulations require that engine manufacturers must reduce emissions of particulate
matter (PM). PM is made up mainly of carbon in the form of PM and complex compound that are absorbed into the PM
particles. Significant quantities PM poses a threat to health. Technologies available for PM reduction are heavily de-
pendent on after-treatment systems, which are respectively active and passive. An active trap system requires a control
unit to trigger and control the regeneration process (clean the trap system). The measurement of PM is crucial for the
trap system to enable the fine judgment as to when to initiate the process. If the regeneration is too infrequent, the filter
will block. The frequency of the regeneration and the life of the filter are compromised. Diesel engine particular matter
prediction has always been a major challenge to the industry. A simple way to handle the PM estimation is to use
black-box modelling as described in [J. Deng, B. Maass, R. Stobart, PM prediction in both steady state and transient
operation of diesel engines, Proc IMechE, Part D: Journal of Automobile Engineering, 2011, 225, in press, DOI:
10.1177/0954407011418029]. This method is used to estimate the PM successfully in both steady and transient engine
operation condition. The main question is how robust and accurate the neural networks are for a regeneration trigger
signal of diesel particular filters. In order to answer such a question, the robust test of PM estimation is carried out
based on different composition of bio-diesel. In this paper, regular diesel fuel will be blended with up to 20% bio-diesel
to test the effect of different fuel resources on particulate matter. Bio-diesel is often added to regular diesel fuel to im-
prove the burning properties and reduce carbon emissions, also is alternative source of fuels. The aim of this paper with
these tests is to ensure that with a realistic change in the fuel composition the estimation of PM is still accurate. There-
fore, neural networks could be used to produce a regeneration signal for diesel particular filter. In this paper a virtual
sensor is proposed to measure the PM. The purpose of the proposed virtual sensor is to estimate the accumulation of PM
and to trigger a regeneration cycle. The virtual sensor based on neural network is used to estimate the PM. In this pa-
per, the performance, robustness and accuracy of simulated the sensor are evaluated in measuring the particulate
amount for non-road transient cycle tests.
Keywords: Neural Networks; Particulate Matter; Smoke; Prediction; Robustness
The increasingly stringent emissions regulations require
that engine manufacturers must reduce emissions of par-
ticulate matter (PM). PM is made up mainly of carbon in
the form of PM and complex compound that are ab-
sorbed into the PM particles. Significant quantities of PM
pose a threat to health. Technologies available for PM
reduction are heavily dependent on after-treatment sys-
tems, which are respectively active and passive. Diesel
Particulate Filters (DPFs) are kinds of after-treatment
system and can effectively reduce the level of PM emis-
sions to ambient background levels .
If PM accumulates on the DPF, the large amount of
heat release during regeneration cannot effectively be
dissipated. This could result in filter damage, for exam-
ple, the formation of cracks or regions which may be
locally melted. Extensive studies try to identify all
mechanisms related to DPF failures and to develop
strategies to avoid these failures . On the other hand, if
the DPF regenerates too frequently, it will also cause
additional fuel penalties. Therefore, accurate knowledge
of the DPF PM loading state is important for improving
fuel economy and extending DPF service life. It is also
very critical for the upcoming on-board diagnostics
A normal DPF system is very expensive (a heavy-duty
catalyzed DPF unit costs about $5000) and a special de-
sign one will cost much more. Therefore, controlling the
Copyright © 2013 SciRes. CN
J. M. DENG ET AL.
DPF regeneration safely is highly critical . The timing
for safely and efficiently regenerating the DPF has be-
come a major industrial challenge.
It can be seen that a control unit to trigger and control
the regeneration process is crucial for both DPF life and
fuel economy. The measurement of PM is crucial for the
trap system to enable the fine judgment as to when to
initiate the process. The frequency of the regeneration
and the life of the filter are compromised.
Diesel engine PM Prediction has always been a major
challenge to the industry [6, 7, 8]. The conventional
method to estimate the PM loading is pressure drop
measurements. But it is affected by exhaust flow varia-
tions and exhibits a low degree of sensitivity to DPF PM
loading and has bad dynamic response over transient
operating conditions. There is study investigating the use
of radio frequency (RF) to directly monitor measure DPF
PM accumulation levels. However, this technique is not
mature enough to be applied in commercial applications
as it is not easy to be calibrated. Hence, more reliable,
stable and accurate PM loading estimation or sensing
method should be studied. Computational fluid dynamics
(CFD) based PM models are computationally intensive
and are not suitable for control purpose and real time
measurement. Recently, neural networks have been used
in a wide variety of automotive applications. The advan-
tage of neural networks is their ability to be used as an
arbitrary function approximation mechanism which has
no requirement to represent the complex underlying
process and is an economic way to obtain the measure-
ment. Model based PM emission prediction method
could be a good alternative way . Neural networks
have been successfully used for emissions prediction .
He et al.  built a model that considers several engine
parameters such as boost pressure and exhaust gas recir-
culation (EGR) and it generates several outputs including
PM emissions. Maass et al. presented a smoke prediction
neural network model using a three-layer autoregressive
model with exogenous inputs (NLARX) model to predict
PM . Bose and Kumar  use fuzzy logic to predict
the engine emissions.
The simple way to handle the PM estimation is to use
black-box modeling as described in .This method is
used to estimate the PM successfully in both steady and
transient engine operation condition. The disadvantage is
the robustness and accuracy of the PM prediction based
on neural networks while different fuels are used in the
engines. The neural network model is built on training
data, which has no information on fuel types. Therefore,
the test of robustness and accuracy is important for the
PM estimation based on neural networks. In order to
carry out the robust and accurate tests, different compo-
sitions of Bio-diesel with standard diesel EN590 are
In order to improve the engine performance and emis-
sions and ensure the sustainability of the fuel supplies,
bio-diesel is a promising alternative for diesels. Bio-
diesel leads to significant reductions in PM emissions,
although the percentage reduction varied with fuel com-
position and engine technology. The average PM reduc-
tions are 26% compared to conventional diesel fuel .
In this paper, standard diesel fuel (EN590) will be
blended with up to 20% bio-diesel fuel to test the effect
of different fuel resources on PM. The aim with these
tests is to ensure that with a realistic change in the fuel
composition the estimation of PM based on neural net-
works is still accurate enough to trig a regeneration cycle.
The virtual sensor of PM measurement based on neural
networks will be used to estimate the PM. With the
simulated sensor we will evaluate its performance, ro-
bustness and accuracy in predicting the PM.
In this paper a virtual sensor is proposed to measure
the PM. The purpose of the proposed virtual sensor is to
estimate the PM and to trigger a regeneration cycle for
diesel particulate filter.
In Section 1 - introduction, a brief background of the
research is introduced. Section 2 reviews the non-linear
autoregressive model with exogenous inputs (NLARX)
neural networks that can be used to predict PM. Section
3 described details of the test facility. Section 4 describes
the data collection and neural network training and ro-
bustness test. Section 5 provides conclusions on this
2. Neural Networks
The field of virtual sensing has become more and more
popular with growing systems complexity such as in
combustion engine control. Its origin lies in the field of
estimators which are specified through physical and nu-
merical relations whereas virtual sensors are character-
ized through black-box approaches such as neural net-
Neural networks can be split into the following three
1) single-layer feed forward networks (SLFN),
2) multi-layer feed forward networks (MLFN),
3) recurrent neural network (RNN).
The chosen network structure or architecture is crucial
for the output performance. Depending on the systems
characteristic: linear or non-linear, static or dynamic, the
network needs to be designed accordingly. Here, the pre-
diction of PM is recognised as highly dynamic and
non-linear that implies a recurrent network structure has
to be chosen to offer sufficient predictive capability. The
NLARX structure can accommodate the dynamics of the
system by feeding previous network outputs back into the
input layer. It also enables the user to define how many
Copyright © 2013 SciRes. CN
J. M. DENG ET AL. 55
previous output and input time steps are required for
representing the systems dynamics best. In this paper a
NLARX model is applied as it is suitable for non-linear-
ity of the problem. Although an important result of ap-
proximation theory is that a three-layer feed-forward
neural network with sigmoid activation functions in the
hidden layer and linear activation functions in the output
layer has the ability to approximate any continuous map-
ping to arbitrary precision, provided that the number of
units in the hidden layer is sufficiently large . How-
ever, the performance of feed-forward neural networks is
limited due to limitations to the number of units in the
hidden states. Performance is further limited by the
memory of personal computers. It is for this reason that,
SLFN and MLFN have not formed part of the work re-
ported in this paper.
A typical structure of an NLARX model is illustrated
in Figure 1. The inputs are represented by u(n) and the
outputs are described by y(n). The formulation of this
NLARX model can be described as:
( )((-1),...,(-),( ),...,(-1))ynFynyn nyunun nu (1)
where ny is number of past output terms used to predict
the current output, nu is number of input terms used to
predict the current output.
Each output of an NLARX model is a function of re-
gressors which are transformations of past inputs and
past outputs. Usually this function has a linear block and
a nonlinear block. The model output is the sum of the
outputs of the two blocks. Typical regressors are simply
delayed input or output variables. More advanced re-
gressors are in the form of arbitrary user-defined func-
tions of delayed input and output variables.
NLARX model training can be cast as a non-linear
unconstrained optimization problem:
min( ,)( )(|)
is a training data set,
y(k) represents the measured output which is the meas-
ured PM in the training set, k |)
is the NLARX
output which is predicted PM, ||.||2 is 2-norm operation,
is a parameter vector, where 1
[, ,, ,]
and p is the number of parameters.
The training process is described as follows. Given a
neural network described by Equation (1), there is an
error metric, that we refer to as performance index of
Equation (2), which is to be minimized. This index is a
representation of the approximation of the network to
some given training patterns. The task will be to modify
the network parameters
to reduce the index
over the complete trajectory to achieve the
minimal value. In this paper the neural networks are
trained using gradient descent algorithms while the initial
is perturbed several times in order to avoid
the local minimal solution. The gradient descent methods
will calculate the vector whose elements are
. The training algorithm will find
the parameters of the network for which the performance
index has reached a desirable value. Given a vectorising
trajectory for the network output and training patterns,
the performance index is the Euclidean norm of the error
matrix of the whole training batch for the output PM.
This model has predicted PM accurately with R-square
= 0.99 as shown in . The inputs for this model are
torque, speed, the first deritatives of torque and speed,
and the second deritative of torque. This paper would
like to investigate how robust the PM sensors are and
whether it is suitable to give an alarm signal to trig the
regeneration cycle for the DPF when different composi-
tion of fuels are used.
NLARX could be used to predict the PM of diesel en-
gines and thus produce a trigger signal for the regenera-
tion cycles of DPFs. This is a virtual sensor (software)
that takes account of various engine parameters to calcu-
late particulate matters.
The advantage of neural networks is their ability to be
used as an arbitrary function approximation mechanism
which has no requirement to represent the complex un-
derlying process and is an economic way to obtain the
measurement. A potential disadvantage of the neural
network is that how robust and accurate it is for the pre-
diction of PM when different fuels are used on engines.
In order to answer these questions, bio-diesel blended
with EN590 will be used for the robust and accurate test
of neural networks bio-diesel is used as an alternative
diesel has become more attractive due to its low emis-
Figure 1. NLARX canonical structure.
Copyright © 2013 SciRes. CN
J. M. DENG ET AL.
3. Test Facility
The engine employed in this study is a Peugeot 2.0 L
HDI Engine. This is a 4-cylinder engine with a Bosch
common rail fuel system and turbo-charger. The engine
calibration used in this work produces a peak torque of
149 Nm at 2002 rpm.
The engine is fully instrumented to measure air, fuel
and cooling system pressures, temperatures and flow
rates. Emissions data is gathered principally using an
AVL 415 smoke and 439 opacity meters (offering steady
state and transient measurement respectively) and a Ho-
riba Horiba MEXA 9100 exhaust gas analyser measuring
nitrous oxide, carbon dioxide, carbon monoxide, unburnt
hydrocarbons and oxygen..
Figure 2 shows the engine facility. The engine is situ-
ated in the lab of the Mechanical and Automotive Eng of
Kingston University. An AVL 439 opacity meter is inte-
grated into the engine exhaust and provides a fast meas-
urement of the exhaust particulate concentration. This
instrument is highly suited to the study of engine
speed-torque transient events during which the control of
exhaust PM is difficult; events such as these contribute
significantly to the total PM produced by the engine.
Accurately predicting the particulates produced during
these events is essential for any model. This is particu-
larly so for a model which has the potential to be de-
ployed as a virtual sensor for determining the optimum
point in time for diesel predicate filter regeneration.
Figure 2. Peugeot 2.0-litre diesel engine and particulate
4. Data Collection
All the engine parameters are recorded in a 100 HZ sam-
pling frequency under the test conditions. For an initial
model set-up NRTC (Nonroad Transient Cycle) are op-
erated on a Peugeot 2.0 L HDI Engine at the test facili-
ties shown in Figure 1. The ECM is set to standard cali-
bration mode. Peugeot 2.0 L HDI Engine runs on a tran-
sient NRTC in order to catch as much as dynamics of the
NRTC is an engine dynamometer transient driving
schedule of total duration of about 1200 seconds. The
speed and torque during the NRTC test is shown in Fig-
ure 3 and Figure 4. Motivation for this choice of cycle is
twofold. First, experience has shown that this is one of
the most challenging cycles in terms of emissions model-
ling. Secondly, meeting emission formation requirements
under the NRTC cycle is also a major concern to engine
manufacturers. The current trend is to design engines
which are marginally passing legislative emission test,
thereby the use of reliable and highly accurate emissions
models is of critical importance.
Figure 3. The torque information of the NRTC.
Figure 4. The speed information of the NRTC.
Copyright © 2013 SciRes. CN
J. M. DENG ET AL. 57
The test is completed with 70% maximum load and
full speed range covering a wide range of engine tran-
sients in different frequencies and combinations. The test
facility is only allowed to run NRTC at 70% load, also
there are some points that the torque drops and this prob-
lem cannot be solved in the current engine setting-up.
Figure 5 the PM results of NRTC
In order to test the robustness of the PM virtual sensor
based on neural network model. Different bio-diesel
blends mixed with standard diesel have been tested for
PM in the engine test cell shown in Figure 2. The blend
is made up of the standard diesel fuel (EN590) and dif-
ferent portion of bio-diesel fuel in volume in the experi-
ment. Different fuel blends which are made of up to 20%
biodiesel-diesel will be used to test the effects of differ-
ent fuel sources on PM emission. The test results are
shown in the Figure 5. It can be seen that the PM with
different percentages of bio-diesel will be different.
Therefore it is necessary to test whether the PM predic-
tion based on neural networks is robust enough to make a
PM estimation for the different bio-diesel blends.
The test results of the standard diesel and the blend of
the standard diesel and 20% bio-diesel have been shown
in Figure 6. It can be seen clearly that the standard diesel
produce higher PM emissions than that of 20% bio-diesel
blend with the same engine and without modification
with hardware and software.
Figure 5. Comparisons of the result for different percentage
bio-diesel and diesel.
Figure 6. Comparison of opacity emission between standard
fuel and 20% bio-diesel during NRTC.
5. Robustness Test of PM Sensors
The PM virtual sensor which is an NLARX model iden-
tified from engine test data is implemented  and
tested against data that collected for the blends of differ-
ent percentages of bio-diesel and standard diesel. In order
to test the robustness of the virtual sensor, the data of the
blend of 10% bio-diesel and standard diesel is used to
train the neural networks. 160 seconds of NRTC data
between 240 - 400 seconds will be enough to train a neu-
ral network . The data between 240 and 400 seconds
data for training purpose is used. The rest of NRTC data
will be used for validation purpose. Inputs for training
the neural network using torque, speed and their deriva-
tives as inputs.
The resulting NLARX model achieves a result of
Rtrain2 = 0.99and Rvalid2 = 0.99 as shown in Figure 7.
This result shows that even if other data, uncorrelated to
the used training data can be predicted to a highly suffi-
cient correlation standard. The inputs are torque, speed
and its derivatives. They provide sufficient feature detail
for the network to generalise unseen and uncorrelated
The PM model based on neural networks has been ob-
tained by using the 10% bio-diesel testing data. The es-
timation of PM based on this model is very accurate for
5% bio-diesel blend (Rpredict2 = 0.99), 10% bio-diesel
blend (Rpredict2 = 0.99), 15% bio-diesel blend (Rpredict2
= 0.99), 20% bio-diesel blend (Rpredict2 = 0.99). This
accuracy can also be seen from the prediction figures in
Figure 7, Figure 9, Figure 10, Figure 11, which show
the PM comparison between the neural network model
and NRTC cycle tests for different percentage blends.
But it cannot predict the standard diesel (Rpredict2 = 0.66)
accurately, shown in Figure 8. It can be concluded that
PM models based on neural networks, which are trained
based on 10% bio-diesel, are quite robust to predict the
emissions for the blends of different percentage
bio-diesels and standard diesel. However, the prediction
for the PM of the standard diesel is not accurate.
Figure 7. NLARX correlation results for training with 160
seconds data and validation agains residual cycle data –
Network Topology: y(n+1) = F(y(n), …, y(n-4), u(n-1)).
Copyright © 2013 SciRes. CN
J. M. DENG ET AL.
Figure 8. PM measured by opacity meter and estimated by
NN model for Standard diesel EN590.
Figure 9. PM measured by opacity meter and estimated by
NN model for 95% EN590 and 5% bio-diesel.
Figure 10. PM measured by opacity meter and estimated by
NN model for 85% EN590 and 15% bio-diesel.
Figure 11. PM measured by opacity meter and estimated by
NN model for 80% EN590 and 20% bio-diesel.
6. Conclusions and Discussion
Neural network model has been used to predict the PM
accurately for the particular blended bio-diesel. For other
percentage blended bio-diesel, the neural networks could
still predict PM accurately. But the prediction is not ac-
curate for the PM of the standard diesel based on 10%
bio-diesel data. For the tests, it could be seen that the
blend of bio-diesel with the standard diesel fuel could
reduce PM emission without any engine modification.
The PM model based on neural networks which is trained
by one of the 10% bio-diesel NRTC transient test is good
for PM estimation with other percentages of bio-diesel,
but has big error with EN590 standard diesel fuel in this
The authors would like to thank the financial support of
Royal Academy of Engineering for the research ex-
change scheme with China.
 A. P. E. York, M. Ahmadinejad, T. C. Watling, A. P.
Walker, J. P. Cox, J. Gast, P. G. Blakeman and R. Al-
lansson, “Modeling of the Catalyzed Continuously Re-
generating DPF (CCR-DPF) System: Model Develop-
ment and Passive Regeneration Studies,” SAE 2007-
 Allansson, R. Cooper, J. Barry, J. E. Thoss, A. Uusimäki,
A. P. Walker and J. P. Warren, “European Experience of
High Mileage Durability of Continuously Regenerating
Diesel Particulate Filter Technology,” SAE 2000-
 R. Allansson, P. G. Blakeman, B. J. Cooper, H. Hess, P. J.
Silcock and A. P. Walker, “Optimising the Low Tem-
perature Performance and Regeneration Efficiency of the
Continuously Regenerating Diesel Particulate Filter
(CR-DPF) System,” SAE 2002-01-0428.
 R. Zhan, Y. Huang and M. Khair, “Methodologies to
Control DPF Uncontrolled Regenerations,” SAE
 M. P. Sturgess, S. F. Benjamin and C. A. Roberts, “Spa-
tial Conversion Profiles within an SCR in a Test Exhaust
System with Injection of Ammonia Gas Modelled in CFD
Using the Porous Medium Approach,” SAE
 Y. Liu, “Diesel Engine Modelling and Optimization for
Emission Reduction,” PhD Thesis, University of Wiscon-
 I. M. Kennedy, “Models of PM Formation and Oxida-
tion,” Progress in Energy and Combustion Science, Vol.
23, No. 2, 1997, pp. 95-132.
 H. P. M. Bockhorn, “Formation in Combustion: Mecha-
nism and Models,” Springer-verlag, Heidelberg, Vol. 59,
 A. Westlund, “Measuring and Predicting Transient Diesel
Engine Emissions,” Academic Thesis, Royal Institute of
Technology, Stockholm, 2009.
 B. Wu, R. G. Prucka, Z. S. Filipi, D. M. Kramer and G. L.
Ohl, “Cam-Phasing Optimization Using Artificial Neural
Networks as Surrogate Models—Fuel Consumption and
NOx Emissions,” SAE 2006-01-1512.
 Y. He and C. J. Rutland, “Application of Artificial Neural
Networks in Engine Modeling,” International Journal of
Engine Research, 2004, Vol. 5, No. 4, pp. 281-296.
Copyright © 2013 SciRes. CN
J. M. DENG ET AL.
Copyright © 2013 SciRes. CN
 B. Maass, R. Stobart and J. Deng, “Diesel Engine Emis-
sions Prediction with Parallel Neural Networks,” Ameri-
can Control Conference, 2009.
 N. Bose and N. S. Kumar, “Prediction of Engine Emis-
sions through Fuzzy Logic Modelling,” International
Conference on Computational Intelligence and Multime-
dia Applications, 2007.
 J. Deng, B. Maass and R. Stobart, “PM Prediction in Both
Steady State and Transient Operation of Diesel Engines,”
Journal of Automobile Engineering, Vol. 226, No. 2,
2011, pp. 260-274. doi:10.1177/0954407011418029
 T. L. Alleman and R. L. McCormick, “Fischer-Tropsch
Diesel Fuels – Properties and Exhaust Emissions: A Lit-
erature Review,” SAE 2003-01-0763.
 K. Funahashi, “On the Approximate Realization of Con-
tinuous Mappings by Neural Networks,” Neural Networks,
Vol. 2, No. 3, 1989, pp. 183-192.
 J. Deng, M. Bastian and R. Stobart, “Minimum Data Re-
quirement for Neural Networks Based on Power Density
Analysis,” IEEE Transactions on Neural Networks, Vol.
23, No. 4, 2012, pp. 585-595.