Professional Documents
Culture Documents
A Study On Extreme Learning Machine For Gasoline E
A Study On Extreme Learning Machine For Gasoline E
fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000.
Digital Object Identifier 10.1109/ACCESS.2020.Doi Number
ABSTRACT This research presents an extreme learning machine (ELM) based neural network modeling
technique for gasoline engine torque prediction. The technique adopts a single-hidden layer feedforward
neural network (SLFN) structure which has the potential to approximate any continuous function with high
accuracy. To verify the robustness of this technique, over 3300 data points collected from a real-world
gasoline engine are used to train, validate, and test the model. These data points cover a wide spectrum of
normal engine operating conditions, with the engine speed from 1000 rpm to 4500 rpm, and the engine torque
from idle to full load. The experiment results demonstrate that the model can predict the gasoline engine
torque with high accuracy. Moreover, this research proposes a weight factor approach to further improve the
prediction accuracy of the model in the desired data regions without modifying the input data set. The
evaluation shows that the weight factor approach can reduce the overall prediction errors in the regions
significantly. This feature is particularly useful in tuning the performance of the model when the significance
of the individual data points varies, or when the distribution of the data points is imbalanced. In practice, the
modeling approaches presented in this research will help reduce the engine test and verification time.
INDEX TERMS Artificial neural network, Extreme learning machine, Gasoline engine model, Torque
prediction, Weight factor
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
engine or its subsystems is required, which has limited the implement, especially for the non-experts, due to the
non-expert’s accessibility to such approaches. Moreover, as disadvantages such as difficulty in determining proper
pointed out in [12], such modeling strategies often fail to work learning rate, getting trapped in local minima, prone to over-
desirably in practice due to the complexity of multivariate training, and very time-consuming for most of the applications
coupling and unexpected noise. [22]. Secondly, these neural network models are typically
In contrast to the above-mentioned modeling approaches, created and tested with about 80 to 130 data points. Thus, there
the data-driven modeling techniques can work well even is still room to explore the robustness of a neural network if
without a deep understanding of the fundamental physics of the number and complexity of the data points have increased
the target system. Generally, these techniques can create significantly. Thirdly, as all the data points are treated equally
models directly from the data collected on the target system in the neural network optimization process, some practical
instead of from the complicated physical/mathematical engineering considerations can be further investigated. For
equations. Moreover, the data-driven modeling techniques instance, as the gasoline engines usually operate at certain
could function well when multivariate correlation exists. conditions much more often than other conditions, a neural
Among many data-driven techniques for the engine-related network model is desirable to have higher accuracy at the
applications, artificial neural network (ANN) is very popular more frequently operated engine conditions, even though the
and has been increasingly studied. For instance, Wu et al. have accuracy at the less frequently operated engine conditions has
proposed a mass air flow (MAF) sensor model on a dual-cam to be sacrificed slightly.
engine using a two hidden layer ANN structure [2]. The Among many ANN training algorithms, neural network
simulated results have shown good agreement between the with random weights (NNRW) is appealing due to its
dynamometer data and the vehicle test data. Togun and advantages in network simplicity, faster training speed, and
Baysec have demonstrated the capability of predicting engine competitive accuracy [23]. Compared to the NNRWs, such as
torque and brake fuel specific consumption using neural Random Vector Functional Link (RVFL) [24] and standard
networks [13]. Cay has investigated the applicability of ANN feed-forward neural network with random weights (the
to predict engine performance and exhaust temperature values Schmidt’s method) [25], ELM outshines in the aspects like
[14]. Cycle-to-cycle combustion variation is studied by Di setting the bias of the output node to zero, transforming
Mauro et al. in [15], which helps identify the pre-ignition and different hidden nodes to one unified form and generating
pre-combustion factors, as well as predict the variation of the random node parameters prior to knowing the training data
indicated mean effective pressure. A neural network based [26]. In addition, ELM can be extended to multiple hidden
nonlinear predictive control scheme has been investigated by layer architecture and has a strong potential for big data
Hu et al., in which the coordinated control of throttle and analytics [23], [27].
wastegate on a turbocharged gasoline engine is explored [16]. In this paper, we present an approach to create a single-
Li et al. have presented a model for nitrogen oxide and smoke hidden layer feedforward neural network based regression
emissions on a diesel engine [17]. In their research, the authors model using ELM that can predict the output torque of a
have not only discussed how to select the model inputs based gasoline engine. The contributions of this paper are threefold:
on the physical analysis, but also showed the insightful 1) It establishes a procedure to create an engine torque
comparisons of a range of neural network architectures in prediction model using ELM approach so that a non
terms of model complexity and accuracy. In addition to engine expert can also benefit from it.
characterizing engine performance and predicting engine 2) It trains and evaluates the neural network with a large
behaviors, neural network models have also been used in amount of experimental data, which not only
engine fault diagnostics. Zheng et al. have revealed how demonstrates the capability of the modeling approach, but
misfire conditions can be detected and categorized under also validates its ability to generalize.
different engine conditions using Elman neural network [18]. 3) It proposes a weight factor approach to further enhance
Bearing knock fault features can be detected by proper ANNs the model accuracy in the desired data regions.
as demonstrated by Chen and Randall [19]. Vibration data are The rest of the paper is organized as follows. Section II
investigated by Ahmed et al. with an ANN model that can introduces the background of the linear regression problem
identify various engine faults and categorize the severity of the using an SLFN and presents an ELM approach to train the
faults [20]. Wen et al. have proposed a bearing fault detection SLFN. It also proposes a weight optimizing algorithm to
model using convolutional neural network, which can enhance the regression accuracy in the selected areas. Section
autonomously learn the unique features of each type of bearing III discusses the experimental setup and data acquisition
fault and identify the defective bearings with its fault type [21]. approach. Section IV presents the model regression and
However, the neural network approaches in the accuracy enhancement results. Section V concludes this paper.
aforementioned research papers still have some room for
improvement. Firstly, the neural networks all use a
backpropagation algorithm to train the weight of each neuron.
In practice, this algorithm may not be easy enough to
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
where 𝑚 is the dimension of the input data, 𝑘 is the B. EXTREME LEARNING MACHINE APPROXIMATION
dimension of the output data, 𝑤𝑖 = [𝑤𝑖1 , 𝑤𝑖2 , … , 𝑤𝑖𝑚 ] is the FOR SLFN
weight vector connecting the 𝑖𝑡ℎ hidden node and the input Since the gradient-based approximation method, such as
neurons, 𝛽𝑖 = [𝛽𝑖1 , 𝛽𝑖2 , … , 𝛽𝑖𝑘 ]𝑇 is the weight vector backpropagation, has several prominent disadvantages as
connecting the 𝑖𝑡ℎ hidden node and the output neurons, and 𝑏𝑖 pointed out in [22], a different algorithm is preferred in this
is the bias of the 𝑖th hidden neuron. A structure of a general research. It is revealed that ELM has superior training speed
SLFN is illustrated in Fig. 1. The term 𝑤 ∙ 𝑥 is the inner and comparable accuracy over many other popular SLFN
product of 𝑤𝑖 and 𝑥𝑗 , and 𝑔 is the activation function of the 𝑖𝑡ℎ training algorithms, such as radial basis function (RBF),
hidden neuron. In practice, 𝑔 can be any nonlinear continuous support vector machine (SVM) and AdaBoost [30]. However,
function, such as sine, sigmoid, hyperbolic tangent, or radial its application on gasoline engine torque prediction has yet to
basis function. The term 𝑔𝑖 (∙) represents the data feature be explored. Therefore, in this paper, an ELM based SLFN
extracted by the 𝑖𝑡ℎ neuron. The compact version of (1) can be gasoline engine torque model is explored and assessed.
written as: With prefixed 𝑤 ̂ and 𝑏̂ (which are usually randomly
𝑌 = 𝐻𝛽 (2) generated numbers between -1 and 1, the approximation
problem stated in (5) can be simplified as:
Hidden Layer
Input Layer
h1
Output Layer
x1
h2 y1
(𝑤11 , 𝑏11 ), (𝑤21 , 𝑏21 ), …, (𝑤𝑁̃ 1, 𝑏𝑁̃ 1)
(𝛽11 , …, 𝛽1𝑘 )
x2 (𝑤12 , 𝑏12 ), (𝑤22 , 𝑏22 ), …, (𝑤𝑁̃ 2, 𝑏𝑁̃ 2) ……
…… h3 (𝛽𝑁̃ 1, …, 𝛽𝑁̃ 𝑘)
(𝑤1𝑚 , 𝑏1𝑚 ), (𝑤2𝑚 , 𝑏2𝑚 ), …, (𝑤𝑁̃ 𝑚, 𝑏𝑁̃ 𝑚)
yk
xm
̃
𝑁
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
Cooling Tower
Environment
Control System
P
Emission Sampling Dyno Control System
Engine
AC Dyno
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
50 27.0
80 29.7
24.3 33.0 29.7
26.7 25.3 24.6
34.6 32.4
29.2 23.518.3 17.8
12.6 15.2
10.5 12.9
20 29.2 32.0
30 28.8 31.6
35.6
Map
MapSize
Size
35.2
27.9
25.8
39.0
36.4
33.9
31.7
26.6
18.1
32.6
25.5
29.4
22.7
60
70
26.1
90
25.2
100
28.7
23.4
27.7
22.5
31.9 28.6
25.7 25.1 24.3
30.8 27.5
24.7
33.9 31.7
24.8 24.1
27.3 21.6
33.1 30.9
25.4 19.7
18.1 17.7
12.7 13.5
18.0 17.6
10.6 11.3
13.5 12.9
11.3 10.8
40 27.9 30.7 34.1 25.5 35.4 30.8 16.5 18.4 16.0 80 24.3 26.7 29.7 24.6 32.4 23.5 17.8 15.2 12.9
50
60
27.0
26.1
29.7
28.7
m
m xxnn
33.0
31.9
25.3
25.1
34.6
33.9
29.2
27.3
18.3
18.1
12.6
12.7
10.5
10.6
90
100
23.4
22.5
25.7
24.7
28.6
27.5
24.3
24.1
31.7
30.9
21.6
19.7
17.7
17.6
13.5
12.9
11.3
10.8
ol
10 27.7 30.4 33.8 42.9 45.0 36.5 38.3 39.5 36.0
ntr
20 29.2 29.4
0 23.9
30 26.3
28.8 29.4
31.6 37.6
35.2 39.5
25.8 31.8
36.4 33.4
31.7 34.5
18.1 31.4
25.5 22.7
Engine Speed (RPM)
VCT Control o d
10 27.7
40 30.4 30.7
27.9 33.8 34.1
42.9 25.5
45.0 35.436.5 30.8
38.3 16.5
39.5 18.4
36.0 16.0
R C uce
°)
500 1000 1500 2000 2500 3000 3500 4000 4500
80 24.3 26.7 29.7 24.6 32.4 23.5 17.8 15.2 12.9
T(
0 23.9 26.3 29.4 37.6 39.5 31.8 33.4 34.5 31.4
90 23.4 25.7 28.6 24.3 31.7 21.6 17.7 13.5 11.3
30.4 Engine Speed
42.9 (RPM)
Pedal Position (%)
VC
10 27.7 33.8 45.0 36.5 38.3 39.5 36.0 100 22.5 24.7 27.5 24.1 30.9 19.7 17.6 12.9 10.8
Pedal Position (%)
500
20 1000 32.0
29.2 1500 35.6
2000 27.9
2500 39.0
3000 3500
33.9 4000
26.6 4500 29.4
32.6
0 23.9
30 26.3
28.8 29.4
31.6 37.6 25.8
35.2 39.5 36.4
31.8 33.4
31.7 34.5
18.1 31.4 22.7
25.5
10 27.7
40 30.4
27.9
Engine
33.8
30.7
Speed
34.1
(RPM)
42.9 25.5
45.0 35.4
36.5 38.3
30.8 39.5
16.5 36.0 16.0
18.4 Engine Speed (RPM)
Pedal Position (%)
500
20 1000
29.2
50 1500 29.7
32.0
27.0 2000 33.0
35.6 2500 25.3
27.9 3000 34.6
39.0 3500
33.9 4000
26.6
29.2 4500
32.6
18.3 29.4 10.5
12.6
500 1000 1500 2000 2500 3000 3500 4000 4500
0 23.9
30 26.3
28.8
60 29.4 28.7
31.6
26.1 37.6 31.9
35.2 39.5 25.1
25.8 31.8 33.9
36.4 33.4
31.7 34.5
18.1
27.3 31.4
25.5
18.1 22.7 10.6
12.7
0 23.9 26.3 29.4 37.6 39.5 31.8 33.4 34.5 31.4
10 27.7
40 30.4
70
27.9 33.8 27.7
25.2
30.7 42.9 30.8
34.1 45.0 24.8
25.5 36.5 33.1
35.4 38.3
30.8 39.5
25.4
16.5 36.0
18.0
18.4 13.5
16.0 11.3 Engine Speed (RPM)
Map Size 10 27.7 30.4 33.8 42.9 45.0 36.5 38.3 39.5 36.0
Pedal Position (%)
20 29.2 32.0
80 35.6 26.7
24.3 27.9 29.7
39.0 24.6
33.9 32.4
26.6 32.6
23.5 29.4
17.8 15.2
10.5 12.9
70 25.2
100 27.7
22.5 30.8 27.5
24.7 24.8 24.1
33.1 30.9
25.4 19.7
18.0 13.5
17.6 11.3
12.9 10.8 70 25.2 27.7 30.8 24.8 33.1 25.4 18.0 13.5 11.3
20
20 29.2
29.2
50 32.0 35.6
32.0 29.7
27.0 27.9
35.6 33.0 39.0
27.9 25.3 33.9
39.0 34.6 26.6
33.9 29.2 32.6
26.6 18.3 29.4
32.6 12.629.4 10.5
80 24.3 26.7 29.7 24.6 32.4 23.5 17.8 15.2 12.9 80 24.3 26.7 29.7 24.6 32.4 23.5 17.8 15.2 12.9
30 28.8 31.6 35.2 25.8 36.4 31.7 18.1 25.5 22.7
T
30 28.8
60 31.6 28.7
26.1 35.2 31.925.8 25.136.4 33.931.7 27.318.1 18.125.5 12.722.7 10.6
90 23.4 25.7 28.6 24.3 31.7 21.6 17.7 13.5 11.3
VC
FIGURE 3. Illustration of the exponential rise of the map size and complexity as the control parameters increase (the table contents are blurred
for proprietary reasons)
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
TABLE II
MODEL INPUT/OUTPUT PARAMETERS AND STATISTICS
Training Data Set Validation Data Set Testing Data Set
Model Parameters I/O
Min Mean Max Min Mean Max Min Mean Max
Engine Torque (Nm) Output 55.3 323.7 671.6 54.6 321.3 635.6 55 326.7 657.4
Engine Speed (rpm) Input 999 2658 4504 999 2728 4504 999 2586 4504
Intake Manifold Pressure (kPa) Input 24.9 71.4 100.3 24.9 70.4 100.3 25 71 100.3
Barometric Pressure (kPa) Input 98.2 99.7 100.9 98.2 99.7 100.9 98.2 99.8 100.9
Intake Air Temperature (degC) Input 17.7 24.2 35.5 19.3 24.2 35.3 19.4 24.2 34.6
VCT (deg) Input 0 25 50 0 24 50 0 23 50
Spark Advance Angle (deg) Input 0 28 55 6.5 28 55 5.75 27.9 55
Lambda (-) Input 0.84 0.96 1 0.84 0.95 1 0.84 0.96 1
Exhaust Gas Temperature (degC) Input 362.5 712.2 938.7 363.1 716.4 929.3 398.3 707.3 934.2
Exhaust Gas Pressure (kPa) Input 1.7 5.6 13.6 1.9 5.7 13.5 2.1 5.5 13.6
Hidden Layer
Input Layer
h1
x1-Raw x1
Output Layer
Normalization
h2
(𝑤11 , 𝑏11 ), (𝑤21 , 𝑏21 ), …, (𝑤𝑁̃ 1, 𝑏𝑁̃ 1)
x2-Raw x2 (𝑤12 , 𝑏12 ), (𝑤22 , 𝑏22 ), …, (𝑤𝑁̃ 2, 𝑏𝑁̃ 2)
Normalization (β1,…, ββ𝑁̃k)T y Restoration y-Raw
…… h3
(𝑤1𝑚 , 𝑏1𝑚 ), (𝑤2𝑚 , 𝑏2𝑚 ), …, (𝑤𝑁̃ 𝑚, 𝑏𝑁̃ 𝑚)
xn-Raw xm
Normalization
̃
𝑁
As pointed out [42], [43], the empirical initialization of the shall be restored to their original scales to get the raw
weights and biases in the random weight networks with non- engineering values. The network structure with data
iterative training algorithms, such as ELM, may not always normalization and restoration is illustrated in Fig. 4. The
lead to the optimal performance. Therefore, tuning of the training data set are used to train the network in order to extract
randomization range is required to achieve desirable the data feature. The validation data set are used to validate the
performance. In addition, since many of the activation trained model and tune the appropriate 𝛽 estimator. The test
functions (such as sine or sigmoid function) have periodic or data set are used to provide unbiased evaluations of the model
asymptotic output ranges, it is preferable that given the once the model is determined.
randomized weights and biases, the output of the neuron To finalize the model structure, the number of hidden
response should avoid such ranges, so the change of the input neurons also needs to be determined. Traditionally the manual
value can be reflected sufficiently on the output. In this trial-and-error approach is used to find an acceptable number
research, the activation function is selected as sigmoid of the hidden nodes. However, this approach is time-
function. All the input/output parameters (such as those listed consuming and cannot always guarantee that the optimal
in Table II) are normalized to [0, 1], where 0 corresponds to number of hidden nodes exists in the experimental range. In
95% of their minimum values and 1 corresponds to 105% of this research, random orthogonal projection based enhanced
their maximum values, respectively. Such normalization can bidirectional ELM (OEB-ELM) [44] is adopted to find the
create some room of redundancy in case the estimated unseen optimal number of hidden neurons. Compared to B-ELM [45]
values fall slightly out of the range of the training values. The and EB-ELM [46] algorithms, this approach leads to better
randomization range of the hidden neuron weights and biases generalization and a more compact network structure. Five
are both selected as [-1, 1] after parameter tuning. For this sets of the orthogonalized initialization matrices are evaluated
application, such settings would make the overall hidden at each odd-hidden node step to search for the appropriate
neuron inputs reside in [-3.6, 2.9], which is a suitable input initialization set that can deliver the minimum residue error. In
range to avoid the asymptotic output range of the activation order to explore how the increase of the number of hidden
function. After the neural network calculation, the output data neurons would affect the performance of the model, sweeps of
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
the number of hidden neurons versus the normalized RMSE the predicted and observed values of the training data set as
are conducted on the training data set and the validation data shown in Fig. 6 (a). The RMSE is 9.18 Nm and the correlation
set. The residue error exit condition is not considered in the coefficient is 0.9983 between the predicted and observed
sweep so the full curve trend can be revealed. As shown in Fig. values of the validation data set as shown in Fig. 6 (b). The
5, when the number of hidden neurons reaches 32, the RMSE is 9.29 Nm and the correlation coefficient is 0.9981
normalized RMSEs of both data sets start to converge, with between the predicted and observed values of the testing data
both values smaller than 0.1. Based on this discovery, the set as shown in Fig. 6 (c). The results have shown good
number of hidden neurons is determined as 32 for this model. generalization across the training, the validation, and the
1.1 testing data sets. Considering the wide torque distribution in
Normalized RMSE Training Data
1.0 Normalized RMSE Validation Data the data sets and practical engineering sense, the results can be
0.9 considered as within acceptable range.
0.8
Normalized RMSE
Torque [Nm]
Torque [Nm]
0 0 0
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700
Torque [Nm] Torque [Nm] Torque [Nm]
(a) (b) (c)
FIGURE 6. Torque model regression results of (a) the training data set, (b) the validation data set, and (c) the testing data set
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
16 16 16
Training_Overall Validation_Overall Testing_Overall
14 Training_Non-LMSO_Points 14 Validation_Non-LMSO_Points 14 Testing_Non-LMSO_Points
Training_LMSO_Points Validation_LMSO_Points Testing_LMSO_Points
12 12 12
RMSE [Nm]
RMSE [Nm]
RMSE [Nm]
10 10 10
8 8 8
6 6 6
4 4 4
2 2 2
0 5 10 15 20 25 30 35 40 45 50 55 60 0 5 10 15 20 25 30 35 40 45 50 55 60 0 5 10 15 20 25 30 35 40 45 50 55 60
Weight Factor Weight Factor Weight Factor
(a) (b) (c)
FIGURE 8. Impact of the change of weight factor on the RMSEs on (a) the training data set, (b) the validation data set, and (c) the testing data set
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
Corr. Coef.
Corr. Coef.
0.998 0.998 0.998
16 16 16
Training Non-Weighted Valid. Non-Weighted Testing Non-Weighted
14 Training Weighted 14 Validation Weighted 14 Testing Weighted
12.3
12.28
12 10.9
10.92 12 12 11.1
11.08 11.0
10.99
9.3 9.3
RMSE [Nm]
RMSE [Nm]
RMSE [Nm]
10 8.9
8.88 9.30 10 10 9.31
7.8
7.77 7.8
7.80 7.8
7.82
8 8 7.3
7.29 8
6.3
6.29
6 5.3
5.29 6 5.3
5.28 6
3.9 4.2
4.23 4.0
4.00
3.86
4 3.1
3.09 3.0
2.96 4 3.2
3.24 3.0 4 2.9
2.88
3.2
3.25
2.7
2.72 2.6
2.58 2.6
2.65 2.96
2.5
2.45 2.8
2.75
2.1
2.10
2 2 1.3
1.34 2
0 0 0
LMSO1 LMSO2 LMSO3 LMSO4 LMSO5 LMSO1 LMSO2 LMSO3 LMSO4 LMSO5 LMSO1 LMSO2 LMSO3 LMSO4 LMSO5
(a) (b) (c)
FIGURE 10. RMSE comparison of the individual LMSO areas on (a) the training data set, (b) the validation data set, and (c) the testing data set
when weight factor = 30
unified as one general approach. Another interesting validation, and the testing data sets. These observations can be
observation is that when 𝑤 continues to grow, the marginal very helpful in improving the regression model. Firstly, the
benefits such as the decreasing RMSE and the increasing inconsistency of enhancement among different LMSO areas
correlation coefficient on the LMSO points start to diminish, indicates that a unified weight factor does not suit all the data
but the side effects like rising RMSE and decreasing regions. The quantitative difference can help determine better-
correlation coefficient on the other points keep expanding. customized weight factors for the individual data regions.
This trait reveals that the optimal 𝑤 should be determined Secondly, the inconsistent behavior among the training, the
based on the trade-off between the two aspects according to validation, and the testing data sets implies the data partition
the specific applications. may be imbalanced across the data sets. Therefore, the feature
In addition to considering all the LMSO points as a whole, learned in the one data set may not suit other data sets in the
it is worth exploring how much improvement each LMSO area same way. The quantitative difference can indicate the
has with the introduction of the weight factor. In order to do distribution discrepancy, which can help adjust the use of the
so, comparisons of the same LMSO areas before and after available data to deliver better generalization of the model.
applying the weight factor are also conducted in this research. In order to visualize the enhancement brought by the weight
As demonstrated in Fig. 10, the introduction of the weight factor on the individual data points, the comparison of the
factor (e.g., 𝑤 = 30) can reduce the RMSE by approximately observed data, non-weighted regression data, and the weighted
2~8 Nm to the LMSO areas in all data sets. Similar patterns regression data are illustrated in Fig. 12. For convenience, only
can be also observed on the correlation coefficients as shown LMSO1 points are presented. As can be seen in the figure, the
in Fig. 11. Generally, a higher correlation coefficient occurs weighted regression points are closer to the observed points
where a smaller RMSE exists, which means that the than the non-weighted regression points (e.g., the weighted
introduction of weight factor does enhance the regression regression data are closer to the 45-degree line than the non-
accuracy of that LMSO area. However, the extent of weighted regression points). This matches the finding in the
enhancement could be different among the training, the above section that the weighted regression data have smaller
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
Corr. Coef.
Corr. Coef.
0.95
Corr. Coef.
FIGURE 11. Correlation coefficient comparison of the individual LMSO areas on (a) the training data set, (b) the validation data set, and (c) the
testing data set when weight factor = 30
Torque_LMSO1 [Nm]
Torque_LMSO1 [Nm]
Torque_LMSO1 [Nm]
95 95 95
90 90 90
85 85 85
80 80 80
80 85 90 95 100 105 80 85 90 95 100 105 80 85 90 95 100 105
Torque_LMSO1 [Nm] Torque_LMSO1 [Nm] Torque_LMSO1 [Nm]
(a) (b) (c)
FIGURE 12. Comparison of regression results of the LMSO1 points in (a) the training data set, (b) the validation data set, and (c) the testing data
set when weight factor = 30
RMSE and higher correlation coefficient than the non- improvement in the desired regions individually, this research
weighted regression data. has also revealed that optimal overall improvement can be
achieved by customizing the weight factor for each desired
V. SUMMARY AND FUTURE WORK region accordingly.
This paper has presented an ANN approach to predict gasoline The modeling and tuning approaches presented in this paper
engine output torque with high accuracy using an ELM based are particularly helpful for rapid engine model creation and
single-hidden layer feedforward neural network. Ten-fold cross verification. In the future, the authors plan to study the
cross-validation is adopted to stabilize the model. Totally 3356 improvement of the ELM algorithm to achieve even higher
real-world engine experiment points have been collected for accuracy.
this model, in which 2685 and 335 experiment points have
been used to train and validate the model, respectively. The REFERENCES
remaining 336 collected data points have been kept unseen by [1] S. Pischinger, “Current and Future Challenges for Automotive
the model and are used to test the model unbiasedly. The Catalysis: Engine Technology Trends and Their Impact,” Top.
Catal., vol. 59, no. 10–12, pp. 834–844, Jul. 2016, doi:
results have demonstrated the model has the capability in 10.1007/s11244-016-0557-3.
predicting the engine torque with high accuracy. The RMSE [2] B. Wu, Z. Filipi, D.N. Assanis, D. M. Kramer, G. L. Ohl, M. J.
between the observed torque and the predicted torque is about Prucka, E. DiValentin, “Using Artificial Neural Networks for
Representing the Air Flow Rate through a 2.4 Liter VVT Engine,”
9 Nm, which is about 2.7% of the mean torque value. In presented at the 2004 Powertrain & Fluid Systems Conference &
addition, this research has also proposed a weight factor Exhibition, Tampa, USA, Oct. 2004, doi: 10.4271/2004-01-3054.
approach to further enhance the model accuracy in the [3] S. Tolou, R. T. Vedula, H. Schock, G. Zhu, Y. Sun, and A. Kotrba,
“Combustion Model for a Homogeneous Turbocharged Gasoline
designated data regions. The experiment has shown that the Direct-Injection Engine,” J. Eng. Gas Turbines Power, vol. 140, no.
RMSE in these regions can be further reduced by 10, p. 102804, Oct. 2018, doi: 10.1115/1.4039813.
approximately 6 Nm. Moreover, by exploring the quantitative
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
[4] N. Togun, S. Baysec, and T. Kara, “Nonlinear modeling and Trans. Ind. Electron., vol. 65, no. 7, pp. 5990–5998, Jul. 2018, doi:
identification of a spark ignition engine torque,” Mech. Syst. Signal 10.1109/TIE.2017.2774777.
Process., vol. 26, pp. 294–304, Jan. 2012, doi: [22] Guang-Bin Huang, Qin-Yu Zhu, and Chee-Kheong Siew, “Extreme
10.1016/j.ymssp.2011.06.010. learning machine: a new learning scheme of feedforward neural
[5] Z. Tan and R. D. Reitz, “An ignition and combustion model based networks,” in 2004 IEEE International Joint Conference on Neural
on the level-set method for spark ignition engine multidimensional Networks (IEEE Cat. No.04CH37541), Budapest, Hungary, 2004,
modeling,” Combust. Flame, vol. 145, no. 1–2, pp. 1–15, Apr. 2006, vol. 2, pp. 985–990, doi: 10.1109/IJCNN.2004.1380068.
doi: 10.1016/j.combustflame.2005.12.007. [23] W. Cao, X. Wang, Z. Ming, and J. Gao, “A review on neural
[6] P. Boudier, S. Henriot, T. Poinsot, and T. Baritaud, “A model for networks with random weights,” Neurocomputing, vol. 275, pp.
turbulent flame ignition and propagation in spark ignition engines,” 278–287, Jan. 2018, doi: 10.1016/j.neucom.2017.08.040.
Symp. Int. Combust., vol. 24, no. 1, pp. 503–510, 1992. [24] B. Igelnik and Yoh-Han Pao, “Stochastic choice of basis functions
[7] C. S. Daw, C. E. A. Finney, J. B. Green, M. B. Kennel, J. F. Thomas, in adaptive function approximation and the functional-link net,”
and F. T. Connolly, “A Simple Model for Cyclic Variations in a IEEE Trans. Neural Netw., vol. 6, no. 6, pp. 1320–1329, Nov. 1995,
Spark-Ignition Engine,” presented at the 1996 SAE International doi: 10.1109/72.471375.
Fall Fuels and Lubricants Meeting and Exhibition, Toronto, [25] W. F. Schmidt, M. A. Kraaijveld, and R. P. W. Duin, “Feedforward
Canada, Oct. 1996, p. 962086, doi: 10.4271/962086. neural networks with random weights.pdf,” in Proceedings., 11th
[8] S. G. Poulos and J. B. Heywood, “The effect of chamber geometry IAPR International Conference on Pattern Recognition., The Hague,
on spark-ignition engine combustion,” SAE Trans., pp. 1106–1129, Netherlands, 1992, vol. Vol II, pp. 1–4.
1983. [26] G.-B. Huang, “An Insight into Extreme Learning Machines:
[9] G. Fontana and E. Galloni, “Variable valve timing for fuel economy Random Neurons, Random Features and Kernels,” Cogn. Comput.,
improvement in a small spark-ignition engine,” Appl. Energy, vol. vol. 6, no. 3, pp. 376–390, Sep. 2014, doi: 10.1007/s12559-014-
86, no. 1, pp. 96–105, Jan. 2009, doi: 9255-2.
10.1016/j.apenergy.2008.04.009. [27] X.-Z. Wang, R. Wang, and C. Xu, “Discovering the Relationship
[10] J. M. Luján, C. Guardiola, B. Pla, and P. Bares, “Estimation of Between Generalization and Uncertainty by Incorporating
trapped mass by in-cylinder pressure resonance in HCCI engines,” Complexity of Classification,” IEEE Trans. Cybern., vol. 48, no. 2,
Mech. Syst. Signal Process., vol. 66–67, pp. 862–874, Jan. 2016, pp. 703–715, Feb. 2018, doi: 10.1109/TCYB.2017.2653223.
doi: 10.1016/j.ymssp.2015.05.016. [28] Guang-Bin Huang and H. A. Babri, “Upper bounds on the number
[11] R. W. Weeks and J. J. Moskwa, “Automotive Engine Modeling for of hidden neurons in feedforward networks with arbitrary bounded
Real-Time Control Using MATLAB/SIMULINK,”SAE nonlinear activation functions,” IEEE Trans. Neural Netw., vol. 9,
Trans.Mater Manuf Process,Vol. 104, p. 950417, Feb. 1995,doi: no. 1, pp. 224–229, Jan. 1998, doi: 10.1109/72.655045.
10.4271/950417. [29] N. J. Guliyev and V. E. Ismailov, “A single hidden layer
[12] G. Jiang, P. Xie, H. He, and J. Yan, “Wind Turbine Fault Detection feedforward network with only one neuron in the hidden layer can
Using a Denoising Autoencoder With Temporal Information,” approximate any univariate function,” Neural Comput., vol. 28, no.
IEEEASME Trans. Mechatron., vol. 23, no. 1, pp. 89–100, Feb. 7, pp. 1289–1304, Jul. 2016, doi: 10.1162/NECO_a_00849.
2018, doi: 10.1109/TMECH.2017.2759301. [30] Guang-Bin Huang, Hongming Zhou, Xiaojian Ding, and Rui Zhang,
[13] N. K. Togun and S. Baysec, “Prediction of torque and specific fuel “Extreme Learning Machine for Regression and Multiclass
consumption of a gasoline engine by using artificial neural Classification,” IEEE Trans. Syst. Man Cybern. Part B Cybern., vol.
networks,” Appl. Energy, vol. 87, no. 1, pp. 349–355, Jan. 2010, doi: 42, no. 2, pp. 513–529, Apr. 2012, doi:
10.1016/j.apenergy.2009.08.016. 10.1109/TSMCB.2011.2168604.
[14] Y. Cay, “Prediction of a gasoline engine performance with artificial [31] A. E. Hoerl and R. W. Kennard, “Ridge Regression: Biased
neural network,” Fuel, vol. 111, pp. 324–331, Sep. 2013, doi: Estimation for Nonorthogonal Problems,” Technometrics, vol. 12,
10.1016/j.fuel.2012.12.040. no. 1, p. 55, Feb. 1970, doi: 10.2307/1267351.
[15] A. Di Mauro, H. Chen, and V. Sick, “Neural network prediction of [32] G.-B. Huang, Q.-Y. Zhu, and C.-K. Siew, “Extreme learning
cycle-to-cycle power variability in a spark-ignited internal machine: Theory and applications,” Neurocomputing, vol. 70, no.
combustion engine,” Proc. Combust. Inst., vol. 37, no. 4, pp. 4937– 1–3, pp. 489–501, Dec. 2006, doi: 10.1016/j.neucom.2005.12.126.
4944, 2019, doi: 10.1016/j.proci.2018.08.058. [33] G. H. Golub, M. Heath, and G. Wahba, “Generalized Cross-
[16] Y. Hu, H. Chen, P. Wang, H. Chen, and L. Ren, “Nonlinear model Validation as a Method for Choosing a Good Ridge Parameter,”
predictive controller design based on learning model for Technometrics, no. 21(2), pp. 215–223, 1979.
turbocharged gasoline engine of passenger vehicle,” Mech. Syst. [34] W. Pan, “Akaike’s Information Criterion in Generalized Estimating
Signal Process., vol. 109, pp. 74–88, Sep. 2018, doi: Equations,” Biometrics, vol. 57, no. 1, pp. 120–125, Mar. 2001, doi:
10.1016/j.ymssp.2018.02.012. 10.1111/j.0006-341X.2001.00120.x.
[17] H. Li, K. Butts, K. Zaseck, D. Liao-McPherson, and I. [35] K. O. Kvålseth, “Cautionary Note About R 2,” Am. Stat., vol. 39,
Kolmanovsky, “Emissions Modeling of a Light-Duty Diesel Engine no. 4, pp. 279–285, 1985.
for Model-Based Control Design Using Multi-Layer Perceptron [36] O. Harel, “The estimation of R 2 and adjusted R 2 in incomplete
Neural Networks,” presented at the WCXTM 17: SAE World data sets using multiple imputation,” J. Appl. Stat., vol. 36, no. 10,
Congress Experience, Detroit, USA ,Mar. 2017, pp. 2017-01–0601, pp. 1109–1118, Oct. 2009, doi: 10.1080/02664760802553000.
doi: 10.4271/2017-01-0601. [37] K. P. Burnham, D. R. Anderson, and K. P. Burnham, Model
[18] T. Zheng, Y. Zhang, Y. Li, and L. Shi, “Real-time combustion selection and multimodel inference: a practical information-
torque estimation and dynamic misfire fault diagnosis in gasoline theoretic approach, 2nd ed. New York: Springer, 2002.
engine,” Mech. Syst. Signal Process., vol. 126, pp. 521–535, Jul. [38] Z. Gao et al., “Drive cycle simulation of high efficiency
2019, doi: 10.1016/j.ymssp.2019.02.048. combustions on fuel economy and exhaust properties in light-duty
[19] J. Chen and R. B. Randall, “Intelligent diagnosis of bearing knock vehicles,” Appl. Energy, vol. 157, pp. 762–776, Nov. 2015, doi:
faults in internal combustion engines using vibration simulation,” 10.1016/j.apenergy.2015.03.070.
Mech. Mach. Theory, vol. 104, pp. 161–176, Oct. 2016, doi: [39] W. Zong, G.-B. Huang, and Y. Chen, “Weighted extreme learning
10.1016/j.mechmachtheory.2016.05.022. machine for imbalance learning,” Neurocomputing, vol. 101, pp.
[20] R. Ahmed, M. El Sayed, S. A. Gadsden, J. Tjong, and S. Habibi, 229–242, Feb. 2013, doi: 10.1016/j.neucom.2012.08.010.
“Automotive Internal-Combustion-Engine Fault Detection and [40] K. Li, X. Kong, Z. Lu, L. Wenyin, and J. Yin, “Boosting weighted
Classification Using Artificial Neural Network Techniques,” IEEE ELM for imbalanced learning,” Neurocomputing, vol. 128, pp. 15–
Trans. Veh. Technol., vol. 64, no. 1, pp. 21–33, Jan. 2015, doi: 21, Mar. 2014, doi: 10.1016/j.neucom.2013.05.051.
10.1109/TVT.2014.2317736. [41] R. Fletcher, Practical methods of optimization, 2nd ed. Chichester,
[21] L. Wen, X. Li, L. Gao, and Y. Zhang, “A New Convolutional Neural West Sussex England: John Wiley & Sons, Ltd, 2013.
Network-Based Data-Driven Fault Diagnosis Method,” IEEE
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2020.3000152, IEEE Access
[42] W. Cao, J. Gao, Z. Ming, and S. Cai, “Some Tricks in Parameter and alternative fuel combustion, dual-fuel applications, clean & efficient
Selection for Extreme Learning Machine,” IOP Conf. Ser. Mater. combustion enabling, and advanced engine development. He has published
Sci. Eng., vol. 261, p. 012002, Oct. 2017, doi: 10.1088/1757- a number of Journal articles and peer-reviewed Conference papers
899X/261/1/012002. worldwide. Dr. Han is a member of the Society of Automotive Engineers,
[43] M. Li and D. Wang, “Insights into randomized algorithms for neural American Society of Mechanical Engineers, and a licensed Professional
networks: Practical issues and common pitfalls,” Inf. Sci., vol. 382– Engineer of Ontario.
383, pp. 170–178, Mar. 2017, doi: 10.1016/j.ins.2016.12.007.
[44] W. Cao, J. Gao, X. Wang, Z. Ming, and S. Cai, “Random
Orthogonal Projection Based Enhanced Bidirectional Extreme Jimi Tjong received the B.Sc., M.Sc., and Ph.D.
Learning Machine,” in Proceedings of ELM 2018, vol. 11, J. Cao, degrees in Mechanical Engineering from the
C. M. Vong, Y. Miche, and A. Lendasse, Eds. Cham: Springer University of Windsor, ON, Canada, in 1980,
International Publishing, 2020, pp. 1–10. 1984, and 1993, respectively. He is currently an
[45] Yimin Yang, Yaonan Wang, and Xiaofang Yuan, “Bidirectional Adjunct Professor at the University of Windsor,
Extreme Learning Machine for Regression Problem and Its University of Toronto, and McMaster University.
Learning Effectiveness,” IEEE Trans. Neural Netw. Learn. Syst., His research interests include production and
vol. 23, no. 9, pp. 1498–1505, Sep. 2012, doi: comprehensive design validation of components
10.1109/TNNLS.2012.2202289. and systems for conventional and electrified
[46] W. Cao, Z. Ming, X. Wang, and S. Cai, “Improved bidirectional vehicles, and optimization of automotive test
extreme learning machine based on enhanced random search,” systems for cost, performance, and compatibility.
Memetic Comput., vol. 11, no. 1, pp. 19–26, Mar. 2019, doi:
10.1007/s12293-017-0238-1.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/.