Available via license: CC BY-SA 4.0
Content may be subject to copyright.
Int. J. Renew. Energy Dev. 2024, 13 (6),1104-1114
| 1104
https://doi.org/10.61435/ijred.2024.604 63
ISSN: 2252-4940/© 2024.The Author(s). Published by CBIORE
Contents list available at CBIORE journal website
International Journal of Renewable Energy Development
Journal homepage: https://ijred.cbiore.id
Adaptive control of plug-in hybrid electric vehicles based on energy
management strategy and dynamic programming algorithm
Yuxin Ge
*
College of Mechanical and Electrical Engineering, Central South University of Forestry and Technology, Changsha, 410004, China
Abstract. This study mainly analyses the fuel consumption of plug-in hybrid vehicles during operation. A new control method for automobiles based
on energy management strategy and dynamic programming algorithm is proposed. The new method plans and analyses the minimum electricity
consumption, and then uses dynamic programming algorithms to analyse this parameter. The research results indicated that the vehicle state was
constantly changing with the variation of SOC value during driving. The energy mobilization of the vehicle was more obvious after adding dynamic
programming strategy. The efficiency of the vehicle was relatively high in driving state 1, with a minimum value of 70%, which was about 20% higher
than in driving state 4. The average fuel consumption in driving state 2 was 1.8L higher than in other driving states. The overall efficiency of
automobiles after incorporating dynamic programming was improved, with a shorter time to reach the lowest efficiency point compared with not
incorporating dynamic programming algorithms. The highest efficiency value was 7.86% higher than that of not incorporating dynamic programming
models. The new control method can reduce energy consumption and improve the energy management and control effect. The study provides a
better research direction for energy management and control of hybrid electric vehicles in the future.
Keywords: Electric vehicles; Consumption; DP algorithm; Energy management; Control; efficiency.
@ The author(s). Published by CBIORE. This is an open access article under the CC BY-SA license
(http://creativecommons.org/licenses/by-sa/4.0/).
Received: 12th July 2024; Revised: 26th Sept 2024; Accepted: 7th Oct 2024; Available online: 16th Oct 2024
1. Introduction
With the acceleration of globalization and industrialization,
energy consumption and environmental pollution issues are
becoming increasingly severe. Carbon emissions in the
transportation sector have caused enormous pressure on the
environment (Jia et al., 2023). Plug in Hybrid Electric Vehicle
(PHEV), as an energy-saving and emission reducing
transportation vehicle, is a new type of vehicle that can work
together through internal combustion engines and electric
motors. This not only effectively reduces energy consumption,
but also reduces vehicle emissions, which is crucial for
achieving energy conservation and emission reduction in
transportation (Song et al., 2020). The energy efficiency and
environmental performance of PHEV largely depend on the
optimization of energy management strategies. The Dynamic
Programming (DP) algorithm has shown great advantages in
optimizing PHEV energy management strategies due to its
global optimal solution characteristics (Peng et al., 2020).
However, DP still faces high computational complexity and a
high demand for information in practical applications, which
limits their application in real-time or near real-time energy
management systems. Therefore, how to effectively apply DP
to the energy management of PHEV has become an important
direction in current PHEV energy management (Kashif et al.,
2021).
Currently, many experts have carried out in-depth research
on the energy management of PHEVs. Sidharthan et al.
*
Corresponding author
Email: geyuxin1997@163.com (Y. Ge)
proposed a novel adaptive intelligent hybrid Energy
Management Strategy (EMS) to optimize the energy
management of hybrid electric tricycles. The new strategy
utilized absolute energy sharing algorithm and fuzzy logic
controller to ensure efficient utilization of power source and
motor power demand. Compared with BEV, the new strategy
could significantly reduce battery peak power, reduce battery
capacity loss, and lower total operating costs, demonstrating
significant advantages in energy management of hybrid electric
vehicles (Sidharthan et al., 2023). The strategy significantly
reduces the peak battery power and decreases the battery
capacity loss. However, the strategy may not have sufficient
global optimisation capability compared to the DP algorithm.
Belkhier et al. proposed a hybrid battery-FCS energy storage
and management system and passive control technology to
improve the power efficiency and response speed of hybrid
electric vehicles. The research results indicated that the
technology could ensure that hybrid electric vehicles obtained
sustained electricity from hybrid energy resources. The
research results indicated that the new method achieved high-
power integration and improved the operating speed of electric
vehicles (Belkhier et al., 2024). The research proposes a hybrid
battery-fuel cell system and passive control techniques that can
improve the power efficiency and responsiveness of hybrid
electric vehicles. However, the method has some problems with
the stability and durability of the fuel cell under different driving
conditions and environments. Yao et al. proposed a novel offline
online hybrid deep reinforcement learning strategy to optimize
Research Article
Y.Ge
Int. J. Renew. Energy Dev 2024, 13(6), 1104-11 14
| 1105
ISSN: 2252-4940/© 2024. The Author(s). Published by CBIORE
the powertrain control strategy of hybrid electric vehicles and
improve fuel economy. The new strategy utilized offline vehicle
data to establish an initial model and explored new control
strategies through online learning algorithms. Compared with
online learning algorithms, the new method could learn faster
and more stably, significantly improving fuel economy (Yao et al.,
2023). The proposed offline-online hybrid deep reinforcement
learning strategy is able to build an initial model from offline
vehicle data. However, its performance may be insufficient in
variable real-world road conditions with high real-time
requirements. Younes et al. proposed a fuzzy logic controller
ground on predetermined motor torque and battery charging
state to optimize the renewable energy management of hybrid
intelligent vehicles. The new controller could adjust energy
consumption while maintaining driving performance. Through
the implementation and evaluation of simulation models for in
vehicle hybrid power systems, the research results showed that
the new management algorithm could effectively reduce
changes in battery charging status, improve power system
efficiency, and perform outstandingly in different driving cycles
and harsh environments (Younes et al., 2023). The controller
proposed in the study is able to flexibly adjust the energy
consumption according to the real-time conditions of the
vehicle. However, the controller relies on predetermined motor
torque and battery status parameters, which may be limited in
practical applications. Cao et al. focused on hybrid electric
vehicles as a reliable choice to improve fuel economy and
reduce emissions. To fully leverage its advantages, energy
management and torque distribution were important directions
for control strategies. A comprehensive evaluation method was
proposed based on relevant literature. The research results
indicated that energy management strategies provided
important references for the development, control, and
optimization of hybrid vehicles (Cao et al., 2023). The
methodology proposed in the study shows the potential to
improve fuel economy and reduce emissions. However, the
study may limit the credibility and replication of the
methodology in practical applications.
In summary, most of the current research on energy
management in automobiles only focuses on power and fuel
consumption management, and only includes one method for
analysis. There are few directions for optimizing energy
management in automobiles. Therefore, a novel method based
on EMS and DP algorithm is designed. The automotive EMS is
used to analyze the energy consumption status at different
stages, fully understanding the energy consumption of the
vehicle during driving. At the same time, DP is added to reduce
fuel consumption during the driving process and improve the
efficiency of the driving process.
2.Method
2.1 Construction of power vehicle model based on energy
management strategy
When PHEV performs power adaptive control, it is
necessary to first conduct energy analysis on the vehicle model
and working system to build a driving model for the vehicle's
operation. The hybrid charging process of PHEVs mainly has
three modes: series, parallel, and series parallel hybrid. In the
series connection, when the vehicle's battery is low, it is
transmitted through the engine to generate electricity. The
parallel structure allows for direct control of the automotive
system through the interaction between the generator and
engine. The hybrid structure combines series and parallel
structures (He et al., 2021). The hybrid structure vehicle model
is shown in Figure 1.
From Figure 1, in the hybrid structure, the engine and
electric motor are connected through mechanical shafts. The
power of the two engines is distributed through a power
distribution system. Secondly, the motor is connected to the
driving system of the vehicle through mechanical shafts, driving
the hybrid vehicle and achieving vehicle operation. The
vehicle's battery will be connected to the inverter through a
connecting wire during charging, and then connected to the
generator and electric motor through the inverter to achieve the
charging. There are five operating modes in the PHEV system,
including pure electric operation, engine operation, hybrid
operation, engine charging for the motor, and regenerative
braking mode.
The pure electric working mode is where only the generator
operates to provide the system power for the PHEV. However,
in this mode, the operation needs to satisfy certain requirements
for the speed of the motor. If the speed requirements cannot be
met, the power drive of the vehicle will not be solely driven by
the motor. The working mode of the generator is the process in
which the engine drives the vehicle separately. In this mode, the
generator and other motors of the vehicle do not work, which
can minimize the consumption. The hybrid working mode is
when the vehicle system rotates at a high speed and both the
engine and generator cannot reach the operating speed of the
motor, and both operate together. In this driving mode, the main
engine is used as the auxiliary motor. When the power of the
Fig 1 Hybrid structure vehicle model
Y.Ge
Int. J. Renew. Energy Dev 2024, 13(6), 1104-11 14
| 1106
ISSN: 2252-4940/© 2024. The Author(s). Published by CBIORE
engine cannot meet the system requirements, the remaining
power is supplied by the motor. The vehicle charging mode is
operated by the engine as the power drive system to charge the
engine when the vehicle's battery is low. The braking mode of a
vehicle is powered by a generator and stores energy. The
vehicle is displayed in Figure 2.
From Figure 2, the vehicle model framework includes
vehicle driving, vehicle control, and vehicle power modules.
The vehicle control module is mainly driven by the driver's
operation. The control module consists of several modules,
including the vehicle's engine, generator, transmission, and
power battery. The driving module is the main signal emitting
module of the vehicle model, which analyzes the parameters of
the decision-maker through a PI controller and transmits the
parameters into the vehicle control system to achieve vehicle
control. The relationship between automotive control
parameters is shown in equation (1) (Rasool et al., 2023).
(1)
In equation (1), represents the parameter value of the vehicle
brake plate. signifies the proportional value of the parameter.
signifies the integral constant of time. represents the
current driving speed. signifies the target driving speed.
The model needs to manage and analyze its energy. Therefore,
the power system is subjected to pattern analysis, which is to
analyze the fuel consumption of the power system. The fuel
consumption of a vehicle is shown in equation (2) (Punyavathi
et al., 2024).
(2)
In equation (2), represents the fuel consumption
efficiency of power reduction. represents the rotational
torque.( ). represents the rotational speed, measured
by ( ). The power of the engine is displayed in
equation (3) (Zhang et al., 2023; Mohammed et al., 2023).
(3)
In equation (3), represents the engine power. To obtain a
more effective power output, the product of torque and speed is
divided by 9550. The 9550 in the formula is obtained through
unit conversion, which is used to convert the product of torque
and speed into a value in kilowatts (kW). The efficiency of the
electric motor is shown in equation (4) (Venkitaraman & Kosuru,
2023; Milbradt et al., 2023).
(4)
In equation (4), represents the efficiency of the electric
motor. represents the rotational moment of the electric
motor. represents the rotational speed of the electric
motor. The working efficiency of the electric motor at this time
consists of two parts. When the rotational moment of the
electric motor is greater than 0, it means that the vehicle is in
the electric motor working mode. At this time, the power is
shown in equation (5) (Gao et al., 2023; Wang et al., 2023).
(5)
In equation (5), represents the working power of the
generator. When the electric motor is working for the engine, its
power magnitude is shown in equation (6) (Usman & Abdullah,
2023; Song et al., 2023, Ma et al., 2024; Shi et al., 2023).
(6)
To simplify the energy management process of
automobiles, the influencing factors of automobiles are
simplified to only consider the impact of the remaining
electricity of the automobile. The remaining power of a vehicle
is shown in equation (7) (Hua et al., 2023; Yang et al., 2024).
(7)
In equation (7), represents the current charging status
of the battery. represents the initial amount of remaining
electricity in the vehicle. represents the battery charge
loading capacity of the vehicle. represents the current of the
battery at time . The EMS for automobiles has two directions:
automobile power utilization and power maintenance (Tian et
al., 2024; Millo Fetal, 2023).
2.2 Control strategy of stage dynamic programming algorithm for
automobiles
DP is an optimization control model for solving multi-stage
operation decisions of automobiles, mainly for planning energy
decision-making strategies of automobiles. By conducting
functional analysis on the minimum nodes in the automotive
phase, the optimal decision-making method for each
operational phase of the vehicle can be obtained. The decision-
making process of the algorithm is shown in Figure 3.
Fig 2 Automotive system structure
Y.Ge
Int. J. Renew. Energy Dev 2024, 13(6), 1104-11 14
| 1107
ISSN: 2252-4940/© 2024. The Author(s). Published by CBIORE
From Figure 3, when the state parameters are input, the
state parameters of the vehicle are first judged. Then, stage
decisions are made through the DP. Aftermaking the decision,
the state is judged before entering the next DP stage. Its state
can be represented by equation (8) (Vignesh et al., 2023; Cao et
al., 2023; Abd-Elhaleem et al., 2023).
(8)
In equation (8), represents the discrete system of the
DP. represents the control method in the system variable.
belongs to a set of variables in system space. When the
algorithm performs control at different stages, the objective
function size of the current stage can be obtained by solving the
state function of the algorithm, as shown in equation (9) (Ruan
et al., 2023; Gnanaprakasam et al., 2023; Chen et al., 2024).
(9)
In equation (9), denotes the cost function starting
from the initial state . denotes the initial state vector of the
system. denotes the immediate cost function at stage
. denotes the total number of time steps. denotes the final
state at state , and denotes the control inputs at state .
denotes the immediate additional cost function at
stage . denotes the additional cost function at state
. denotes the additional cost function at state
. represents the instantaneous cost
and punishment level at the end of the stage.
represents the total cost at that time. If the driving
status is already known at this point, then, the vehicle status is
controlled to obtain more suitable parameters and kinetic
energy. The energy control equation of the vehicle at this time
is shown in equation (10) (Cui et al., 2024; Hao et al., 2023).
(10)
In equation (10), represents the remaining power
of the transmission. represents the rotational moment of the
generator. represents the variable size that the
algorithm system can manipulate. signifies the system
state. signifies its control variable. At this point, the state
of SOC is shown in equation (11) (Gao et al., 2024; Vignesh &
Ashok, 2023).
(11)
In equation (11), signifies the open circuit voltage of the
battery. signifies the battery resistance. represents its
charge capacity. represents the electrical power of the
battery. Therefore, for the stage DP of automobiles, it is
necessary to first discretize the remaining driving power of the
vehicle, as shown in equation (12) (He et al., 2024).
(12)
In equation (12),
represents the -th state of stage
. represents the discrete processing state of remaining
electricity. After completing the discrete processing, the optimal
parameters for each charging stage are solved to obtain the
optimal control cost for each stage. Finally, the control state
parameters of the vehicle during the operation phase are
obtained by calculating the state. After continuously repeating
this process, the DP for the automotive phase is completed. DP
can provide energy control optimization for the entire stage
(Jung et al., 2024). However, there is also a prerequisite for using
Fig 3 Algorithm decision process
Fig 4 Driving state process
Y.Ge
Int. J. Renew. Energy Dev 2024, 13(6), 1104-11 14
| 1108
ISSN: 2252-4940/© 2024. The Author(s). Published by CBIORE
this optimization method, which is to determine the driving
state. Due to the uncertainty of the driving status, there is
uncertainty in the DP of automobiles, making it more difficult
for DP algorithms to control and manage energy strategies.
Therefore, the driving status of automobiles need to be
recognized (Cipek et al., 2023). The driving process is shown in
Figure 4.
From Figure 4, the driving state requires first extracting the
state data of the vehicle, followed by calculating the feature
parameters of the current driving state. By mapping the
parameter data to the main space, the spatial driving state is
obtained. The approximate value of the driving state is
calculated, and the driving state is classified and recognized.
Finally, the characteristic state data of the driving state are
output. Therefore, there are two calculation methods for
identifying the driving status of a vehicle. One approach is to
analyze past data on the driving status of a vehicle and obtain
characteristic parameters of its historical state. Another way is
to map its parameters through past driving states, obtain
different parameter components, and finally calculate similar
values to obtain the driving state through similar values (He et
al., 2024). The driving state recognition process is shown in
Figure 5.
From Figure 5, among the two methods, the parameter
clustering state obtained by the first method can be transmitted
into the principal component of the second method to connect
the two methods. The second method can continuously judge
the parameters by clustering analysis, and finally obtain the
optimal parameters of the vehicle's driving state. This means
that the second method has a better data parameter processing
method compared to the first method. Finally, if the vehicle is
powered by an electric motor throughout its journey, there is no
Extract data
Characteristic
parameter
Composed of four
components
Driving status data
Operating
condition type
Real data
Characteristic
parameter
Composed of four
components
Cluster center
Method 1 Method 1
Fig 5 Driving state recognition process
Start
Obtaining
driving status
information
Do you need to plan？
Stage
judgment
Dynamic
trajectory
planning Is it within a reasonable range？
Electricity
distribution
Dynamic
programming
End
Y
N
Y
N
Fig 6 Special state process
Y.Ge
Int. J. Renew. Energy Dev 2024, 13(6), 1104-11 14
| 1109
ISSN: 2252-4940/© 2024. The Author(s). Published by CBIORE
need to plan the vehicle's energy throughout the entire process.
When the driving state of a vehicle exceeds its total mileage, it
is also necessary to plan the energy of the vehicle. The process
of this special state is shown in Figure 6.
From Figure 6, during the planning process, firstly, the
driving status information of the vehicle is obtained to
determine whether energy planning for the vehicle is currently
required. If so, the driving status and distance of the vehicle are
evaluated in stages. Then, the dynamic trajectory planning is
performed again. Secondly, it is necessary to determine whether
the energy allocation in the current state is within a reasonable
range. If so, the electricity in each stage is allocated. Then, the
DP is performed again. If not, the driving status planning will be
directly carried out. If electricity planning is required, the
process will be ended directly.
3. Result and discussion
3.1 Results analysis of automotive energy management strategies
To prove the feasibility of the EMS, the MATLAB software
is used to simulate and analyze the EMS. The initial value of
remaining electricity is 0.95. The limit value of remaining
electricity during energy phase switching is 0.30. The common
NEDC driving state is selected as the driving state. Strategies
based on DP, Deep Reinforcement Learning (DRL) (He et al.,
2024) and Model Predictive Control (MPC) (Jung et al., 2024)
combined with dynamic planning are compared, respectively.
He H et al. found that traditional energy management and
control methods for electric vehicles were hindered by
technological bottlenecks, resulting in poor real-time
generalization ability of control strategies. Therefore, the DRL
method was used for control. The results indicated that this
method effectively improvedthe traditional generalization
ability (He et al., 2024). Jung et al. proposed a tram energy
control strategy based on MPC to address the limitations of
traditional tram energy control methods and enhance the
practical application of tram energy control strategies. The new
method reducedthe energy consumption of vehicles and
improved the energy control effect of trams (Jung et al., 2024).
It can be seen that the traditional tram energy control method
cannot control the tram energy well, so it is necessary to use
more advanced control strategy for control. The remaining
power consumption process and energy strategy planning
during the driving state of the vehicle are shown in Figure 7.
From Figure 7 (a), the SOC value of the vehicle decreased
with time during driving. When the remaining battery was high,
the battery decreased significantly. When the driving time
reached around 5500s, the SOC value reached the set minimum
value. At this time, the SOC value began to fluctuate. When the
SOC value dropped to the set threshold, the vehicle began to
enter the energy storage state. When the SOC value exceeded
the set threshold, the vehicle engaged in a hybrid braking state.
From Figure 7 (b), during the driving, when the rotational
moment of the vehicle was large, the vehicle was mainly
powered by the generator. Therefore, the rotational moment of
the vehicle at this time remained high, and the entire variation
was between -50 Nm and 100 Nm. When the driving time
reached 1000s, the rotational moment of the vehicle was larger.
The power supply of the vehicle was completed jointly by the
generator and the engine. This indicates that as the wheelbase
increases, the vehicle begins to operate in an engine powered
state. Compared with Figure 7 (a), at around 5500s, the SOC
value of the vehicle decreased. At this time, the vehicle's energy
was mainly supplied by the engine, so the engine's rotational
torque started to be at a high value and lasted for a relatively
long time. The driving process conforms to the EMS process.
The vehicle data in different energy management strategies is
displayed in Figure 8.
From Figure 8 (a), the driving speed of the vehicle was the
same under different strategies, so the energy use and
mobilization of the vehicle were the same. From Figure 8 (b),
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
01000 2000 3000 4000 5000 6000 7000
SOC change value
SOC minimum limit
Car travel time(s)
SOC
-150
-100
-50
0
50
100
150
01000 2000 3000 4000 5000 6000 7000
Car travel time(s)
Automobile rotational shaft moment(N*m)
The generator provides power
(a) Trend of SOC value variation
(b) Changes in the rotational axis moment of a car
The engine provides power
Fig 7 Remaining energy consumption process and energy strategy planning of automobile driving state
Y.Ge
Int. J. Renew. Energy Dev 2024, 13(6), 1104-11 14
| 1110
ISSN: 2252-4940/© 2024. The Author(s). Published by CBIORE
different energy management strategies had the same change
trend, gradually decreasing with time. When the set threshold
was reached, only the DP-based EMS decreased in magnitude.
This shows that the strategy used in the study can better adjust
the vehicle state after reaching the threshold. This method can
better manage and analyze the energy of vehicles. From Figure
8 (c), when the driving time of the vehicle reached 700s, the
vehicle began to enter fuel consumption. At this stage, it
indicates that the electric motor power of the vehicle enters a
phased charging stage, and the fuel consumption begins to show
an increasing trend with time. However, the fuel consumption
of the research method is relatively low, and the consumption
time is also relatively late. This indicates that the energy
management method used in the study has better management
strategies.
3.2 Simulation results analysis of automotive dynamic programming
algorithm
To verify the effectiveness of vehicle energy management in
various strategies, the initial goal of SOC is set to 0.95, and the
threshold is set to 0.3. The driving state of the vehicle is also
NEDC. Based on the comparison of parameters under different
driving states, the driving state parameters are displayed in
Table 1. Ave signifies the average value, std signifies the
standard deviation, max represents the maximum value, dec
represents the deceleration stage, acc represents the
acceleration stage, idle represents the deceleration stage vehicle
speed, and uni represents the rated power.
From Table 1, in the first state of vehicle driving, the vehicle
may be driving at a moderate speed, with a stable driving speed
and relatively average deceleration and acceleration time. This
situation generally belongs to urban driving sections. In the
second vehicle driving state parameter analysis, the vehicle's
driving speed is relatively fast, and the acceleration and
deceleration time is relatively short. The vehicle speed is the
highest among the three driving states, with 40.5246 km/h,
indicating that this state may be high-speed driving. The
average speed of the third state is low, with 13.4588 km/h, and
the acceleration and deceleration time is relatively short. The
speed of the driving state is not significantly different, indicating
60
00500 1000 1500 2000 2500 3000
Speed(Km*h)
4
00500 1000 1500 2000 2500 3000
Oil consumption(L)
0.4
00500 1000 1500 2000 2500 3000
SOC
Car travel time (s)
Car travel time (s)
Car travel time (s)
(a) Changes in car driving speed under different strategies
(b) Changes in Automotive Rotation Matrix under Different Strategies
(c) Changes in Automotive SOC under Different Strategies
0.2
2
Dynamic programming
DRL MPC
Dynamic programming
DRL MPC
Fig 8 Comparison of vehicle parameters with different energy management strategies
Table 1
Different driving state parameters
Vehicle driving parameters
NEDC 1
NEDC 2
NEDC 3
NEDC 4
Vave
16.9076
40.5246
13.4588
3.6986
uave
24.0884
46.9208
15.8051
6.9774
Vstd
15.9368
25.9348
10.5941
5.2648
Vmax
55.7117
89.1258
39.9595
18.3034
aave
0.4066
0.3458
0.3684
0.2935
adec
-0.4563
-0.4485
-0.3660
-0.2839
amax
0.8791
1.2238
0.6984
0.4898
amin
-1.1236
-1.6316
-0.8712
-0.5835
astd
0.1912
0.2113
0.1475
0.0885
P(tidle)
26.5996
15.7632
14.6452
46.9917
P(tacc)
26.5996
24.6142
29.9457
13.9845
P(tdec)
24.0577
22.2183
29.4171
15.0415
P(tuni)
23.0321
38.3839
25.6024
23.9627
Y.Ge
Int. J. Renew. Energy Dev 2024, 13(6), 1104-11 14
| 1111
ISSN: 2252-4940/© 2024. The Author(s). Published by CBIORE
that the vehicle travels more evenly and the driving speed is not
fast in this state. The average speed in the fourth driving state is
3.6986 km/h. In this state, the driving speed is not fast, which
m ay be the driving state when the roads in the city are relatively
congested. Figure 9 displays the DP image of the vehicle in four
driving states.
From Figure 9 (a), during the driving process, the speed
change did not gradually change with the increase of time, but
there was a phased change. The speed may remain unchanged
with the increase of time and may be in an upward or downward
phase. This may be caused by different road conditions during
the driving, such as acceleration, urban road driving, rural road
driving, etc. From Figure 9 (b), the SOC value changed
differently when the vehicle was in different driving states. In
the first, third, and fourth vehicle conditions, the driving state
showed a decreasing trend over time. However, the SOC value
of the second vehicle condition first decreased to 0.6, then
showed an upward trend, rising to 0.75, and finally decreased
again. This may be because during this process, the vehicle first
drives at low speed, then starts driving at high speed, and finally
drives at low speed again. The EMS can manage the driving of
vehicles in different states. Figure 10 displays the comparison of
vehicle efficiency and fuel consumption under different states.
From Figure 10 (a), in the three driving states of 1, 3, and
4, the utilization efficiency of the vehicle decreased with the
increase of SOC value, with the lowest being driving state 4,
which reached about 60%. Driving state 1 had higher efficiency,
with a minimum value of 70%, which was about 20% higher than
driving state 4. This may be due to the more reasonable energy
planning in driving state 1. In driving state 2, the efficiency
showed an upward trend, which may be due to the acceleration
of the vehicle during this stage, causing an overall efficiency
improvement. The lowest efficiency point was at 0.4SOC, which
was 60%. From Figure 10 (b), the fuel consumption of the
vehicle increased with the consumption of SOC, and its trend
was the same in driving states 1, 3, and 4. The overall fuel
consumption was about 3.4L per 100 Km. The fuel consumption
per 100 Km in driving state 2 showed a significant upward trend
at a 0.5SOC value, which may be due to the acceleration of this
section. The average fuel consumption was about 5.2L, which
was 1.8L higher than other driving states. The error comparison
before and after using the DP model is shown in Table 2.
140
120
100
80
60
40
20
0
Car driving speed(km*h)
0100 200 300 400 500 600
Car travel time (s)
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
02000 4000 6000 8000 10000
1.0
SOC
Car travel time (s)
NEDC 1
NEDC 2 NEDC 3
NEDC 4
(a) Vehicle driving process status
(b) SOC changes during different car driving processes
Fig 9 The DP image of the vehicle in four driving states
50
55
60
65
70
75
80
85
90
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Automobile engine utilization efficiency (%)
SOC consumption per kilometer
0
1
2
3
4
5
6
7
8
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
SOC consumption per kilometer
Average gasoline consumption per 100
kilometers (L)
NEDC 1
NEDC 2 NEDC 3
NEDC 4 NEDC 1
NEDC 2 NEDC 3
NEDC 4
(a) Comparison of car utilization rates in
different vehicle conditions (b) Comparison of fuel consumption in
different vehicle conditions
Fig 10 Comparison of vehicle efficiency and fuel consumption in different driving states
Y.Ge
Int. J. Renew. Energy Dev 2024, 13(6), 1104-11 14
| 1112
ISSN: 2252-4940/© 2024. The Author(s). Published by CBIORE
From Table 2, the overall efficiency of the vehicle was
lower when the DP model was not used, and the trend gradually
decreased with time, reaching the lowest value of 29.35% at
6265s. The efficiency change of the vehicle after adding the DP
model also decreased with the increase of time, but the overall
decrease value was smaller. At the same time, when reaching
the lowest value, the time was 6675s. It was longer than before
adding the strategy, with an efficiency of 29.98%, which was
higher than the model without the strategy. The maximum
efficiency difference between the model before and after adding
was 7.86%, which may be due to the vehicle starting to
decelerate during this period, resulting in the main movement
being the motor movement. Jia et al discussed adaptive model
predictive control in the context of fuel cell hybrid electric
vehicles, providing new ideas for advanced energy
management strategies beyond DP algorithms (Jia et al., 2023).
Peng et al. analyzed the efficiency of hybrid switches and
inverters, which enhanced the analysis of vehicle efficiency and
energy consumption, especially the effectiveness analysis with
DP algorithm (Peng et al., 2020). Venkitaraman and Kosuru
introduced a hybrid deep learning approach to managing
electric vehicle charging, which provided a comparative
perspective on DP-based energy management strategies
(Venkitaraman & Kosuru, 2023). This indicates that different
studies can provide better research ideas for proposing new
methods and better technical support for current research.
4. Conclusion
The research mainly focused on the insufficient adaptability of
energy management strategies and high energy consumption
during the driving process of hybrid electric vehicles. Therefore,
a vehicle energy automatic adaptation method based on EMS
and DP was proposed. The study first analyzed the EMS. Then,
the DP was usedto optimize the model based on the EMS. The
research results indicated that after incorporating the DP
strategy, the energy mobilization of vehicles was more
apparent. The average speed in several driving states was
16.9076 km/h, 40.5246 km/h, 13.4588 km/h, and 3.6986 km/h,
respectively. Driving state 1 had a higher efficiency, with a
minimum value of 70%, which was about 20% higher than
driving state 4. In driving states 1, 3, and 4, the change was the
same, and the overall fuel consumption was around 3.4L per 100
kilometers. The average fuel consumption in driving state 2 was
about 5.2L, which was 1.8L higher than other driving states. The
overall efficiency of automobiles after incorporating DP was
improved, with a shorter time to reach the lowest efficiency
compared with not incorporating the DP model. The highest
efficiency was 7.86% higher than not incorporating the DP
model. From this, the EMS can effectively manage the energy
mobilization of vehicles. After incorporating the DP model,
vehicle energy consumption can be effectively controlled.
Although the research has achieved many results, there are still
some shortcomings. For example, the study only analyzes the
dynamic model of the vehicle, which cannot obtain the real state
of the vehicle. Therefore, it is necessary to further analyze real
vehicle data. At the same time, the study only analyzes several
driving states. Therefore, further analysis will be conducted on
more driving states in the future.
Data availability statement
Data is contained within the article.
Reference
Abd-Elhaleem, S., Shoeib, W., Sobaih, A.A. (2023). A new power
management strategy for plug-in hybrid electric vehicles based on
an intelligent controller integrated with CIGPSO algorithm. Energy,
265(11), 126153-126154.
https://doi.org/10.1016/j.energy.2022.126153
Belkhier, Y., Oubelaid, A., & Shaw, R.N. (2024). Hybrid power
management and control of fuel cells‐battery energy storage
system in hybrid electric vehicle under three different modes.
Energy Storage, 6(1), 511-512. https://doi.org/10.1002/est2.511
Cao, Y., Yao, M., & Sun, X. (2023). An overview of modelling and energy
management strategies for hybrid electric vehicles. Applied Sciences,
13(10), 5947-5948. https://doi.org/10.3390/app13105947
Cao, Y., Yao, M., & Sun, X. (2023). An overview of modelling and energy
management strategies for hybrid electric vehicles. Applied Sciences,
13(10), 5947-5948. https://doi.org/10.3390/app13105947
Chen, C., Wang, X., Lei, Z., & Shangguan, C. (2024). Research on Plug-
in Hybrid Electric Vehicle (PHEV) energy management strategy
with dynamic planning considering engine start/stop. World
Electric Vehicle Journal. 15(8), 350-351.
https://doi.org/10.3390/wevj15080350
Cipek, M., Pavković, D., & Kljaić, Z. (2023). Optimized energy
management control of a hybrid electric locomotive. Machines,
11(6), 589-591. https://doi.org/10.3390/machines11060589
Cui, W., Cui, N., Li, T., Du, Y., & Zhang, C. (2024). Multi-objective
hierarchical energy management for connected plug-in hybrid
electric vehicle with cyber–physical interaction. Applied Energy,
360(4), 122816-122817.
https://doi.org/10.1016/j.apenergy.2024.122816
Gao, H., Yin, B., Pei, Y., Gu, H, Xu, S., & Dong, F. (2024). An energy
management strategy for fuel cell hybrid electric vehicle based on
a real-time model predictive control and pontryagin’s maximum
Table 2
Error comparison of different models
Travel time
SOC efficiency (%)
Efficiency of dynamic programming algorithm (%)
Efficiency deviation value (%)
465s
78.45
79.02
0.57%
832s
76.42
77.53
1.11%
2562s
60.24
61.25
1.01%
3354s
53.59
57.65
4.06%
4195s
50.36
51.26
0.90%
4968s
42.36
43.84
1.48%
5354s
39.21
39.62
0.41%
6265s
29.35
38.31
7.86%
6675s
29.35
29.98
0.63%
Y.Ge
Int. J. Renew. Energy Dev 2024, 13(6), 1104-11 14
| 1113
ISSN: 2252-4940/© 2024. The Author(s). Published by CBIORE
principle. International Journal of Green Energy. 11(4), 1-3.
https://doi.org/10.1080/15435075.2024.2322973
Gao, K., Luo, P., Xie, J, Chen, B., Wu, Y., & Du, R. (2023). Energy
management of plug-in hybrid electric vehicles based on speed
prediction fused driving intention and LIDAR. Energy, 284(10),
1285-1286. https://doi.org/10.1016/j.energy.2023.128535
Gnanaprakasam, C.N., Meena, S., & Devi, M.N. (2023).
Shanmugasundaram N, Sridharan S. Robust energy management
technique for plug-in hybrid electric vehicle with traffic condition
identification. Applied Soft Computing. 133(5), 109937-109938.
https://doi.org/10.1016/j.asoc.2022.109937
Hao, J., Ruan, S., & Wang, W. (2023). Model predictive control based
energy management strategy of series hybrid electric vehicles
considering driving pattern recognition. Electronics, 12(6), 1418-
1419. https://doi.org/10.3390/electronics12061418
He, H., Meng, X., Wang, Y., Khajepour, A., An, X., Wang, R., & Sun, F.
(2024). Deep reinforcement learning based energy management
strategies for electrified vehicles: Recent advances and
perspectives. Renewable and Sustainable Energy Reviews. 192(1),
114248-114249. https://doi.org/10.1016/j.rser.2023.114248
He, L., Chen, F., Tian, P., & Gou, H. (2024). An improved energy
management strategy for hybrid electric powered aircraft based on
deep reinforcement learning. Aerospace Science and Technology,
149(6), 109137-109138.
https://doi.org/10.1016/j.ast.2024.109137
He, R., Yan, Y., & Hu, D. (2021). Optimised adaptive control
methodology for mode transition of hybrid electric vehicle based
on the dynamic characteristics analysis. Vehicle System Dynamics,
59(8), 1282-1303.
https://doi.org/10.1080/00423114.2020.1752923
Hua, M., Zhang, C., Zhang, F., Li, Z., Yu, X., Xu, H., & Zhou, Q. (2023).
Energy management of multi-mode plug-in hybrid electric vehicle
using multi-agent deep reinforcement learning. Applied Energy,
348(10), 121526-121526.
https://doi.org/10.1016/j.apenergy.2023.121526
Jia, C., Qiao, W., Cui, J., & Qu, L.Y. (2023). Adaptive model-predictive-
control-based real-time energy management of fuel cell hybrid
electric vehicles. IEEE Transactions on Power Electronics, 38(2),
2681-2694. https://doi.org/10.1109/TPEL.2022.3214782
Jung, J., Kim, D., Yang, L., & Kim, N. (2024). Optimal energy
management strategy for repeat path operating fuel cell hybrid
tram. Energies. 17(7), 1560-1561.
https://doi.org/10.3390/en17071560
Kashif, M., Singh, B., & Murshid, S. (2021). Solar PV array fed self-
sensing control of PMSM drive with robust adaptive hybrid SOGI
based flux observer for water pumping. IEEE Transactions on
Industrial Electronics. 68(8), 6962-6972.
https://doi.org/10.1109/TIE.2020.3003656
Ma, Z., Luan, Y., Zhang, F., & Coskun, S., (2024). A data-driven energy
management strategy for plug-in hybrid electric buses considering
vehicle mass uncertainty. Journal of Energy Storage, 77(6), 109963-
109964. https://doi.org/10.1016/j.est.2023.109963
Milbradt, D.M., de Oliveira Evald, P.J., Hollweg, G.V., & Gründling, H.A.
(2023). A hybrid robust adaptive sliding mode controller for
partially modelled systems: Discrete-time lyapunov stability
analysis and application. Nonlinear Analysis: Hybrid Systems. 48(1),
101333-101334. https://doi.org/10.1016/j.nahs.2023.101333
Millo, F., Rolando, L., Tresca, L., & Pulvirenti, L. (2023). Development
of a neural network-based energy management system for a plug-
in hybrid electric vehicle. Transportation Engineering, 11(1), 100156-
100158. https://doi.org/10.1016/j.treng.2022.100156
Mohammed, A.S., Atnaw, S.M., Salau, A.O., & Eneh, J.N. (2023). Review
of optimal sizing and power management strategies for fuel
cell/battery/super capacitor hybrid electric vehicles. Energy
Reports, 9(9), 2213-2228.
https://doi.org/10.1016/j.egyr.2023.01.042
Peng, Z., Wang, J., Liu, Z., & Li, Z. J. (2020). Adaptive gate delay-time
control of Si/SiC hybrid switch for efficiency improvement in
inverters. IEEE Transactions on Power Electronics, 2020, 3437-3449.
https://doi.org/10.1109/TPEL.2020.3015803
Punyavathi, R., Pandian, A., Singh, A.R., Bajaj, M., Tuka, M.B., & Blazek,
V. (2024). Sustainable power management in light electric vehicles
with hybrid energy storage and machine learning control. Scientific
Reports, 14(1), 5661-5662. https://doi.org/10.1038/s41598-024-
55988-5
Rasool, M., Khan, M.A., & Zou, R. (2023). A comprehensive analysis of
online and offline energy management approaches for optimal
performance of fuel cell hybrid electric vehicles. Energies, 16(8),
3325-3326. https://doi.org/10.3390/en16083325
Ruan, J., Wu, C., Liang, Z., Liu, K., Li, B., Li, W., & Li, T. (2023). The
application of machine learning-based energy management
strategy in a multi-mode plug-in hybrid electric vehicle, part II:
Deep deterministic policy gradient algorithm design for electric
mode. Energy. 269(6), 126792-126793.
https://doi.org/10.1016/j.energy.2023.126792
Shi, D., Li, S., Liu, K., Xu, Y., Wang, Y., & Guo, C., (2023). Adaptive
energy management strategy for plug-in hybrid electric vehicles
based on intelligent recognition of driving cycle. Energy Exploration
& Exploitation, 41(1), 246-272.
https://doi.org/10.1177/01445987221111488
Sidharthan, V.P., Kashyap, Y., & Kosmopoulos, P. (2023). Adaptive-
energy-sharing-based energy management strategy of hybrid
sources in electric vehicles. Energies, 16(3), 1214-1215.
https://doi.org/10.3390/en16031214
Song, D., Bi, D., Zeng, X., & Wang, S. (2023). Energy management
strategy of plug-in hybrid electric vehicles considering thermal
characteristics. International Journal of Automotive Technology, 24(3),
655-668. https://doi.org/10.1007/s12239-023-0055-0
Song, S., Han, C., Lee, G., McCann, R. A., & Jang, G. (2020). Voltage-
sensitivity-approach-based adaptive droop control strategy of
hybrid STATCOM. IEEE Transactions on Power Systems, 36(1), 389-
401. https://doi.org/10.1109/TPWRS.2020.3003582
Tian, S., Zheng, Q., Wang, W., & Zhang, Q. (2024). Integrated real-time
optimal energy management strategy for plug-in hybrid electric
vehicles based on rule-based strategy and AECMS. International
Journal of Vehicle Design, 94(1-2), 150-175.
https://doi.org/10.1504/IJVD.2024.136239
Usman, A. M., & Abdullah, M. K. (2023). An assessment of building
energy consumption characteristics using analytical energy and
carbon footprint assessment model. Green and Low-Carbon
Economy, 1(1), 28-40.
https://doi.org/10.47852/bonviewGLCE3202545
Venkitaraman, A.K., & Kosuru, V.S. (2023). Hybrid deep learning
mechanism for charging control and management of Electric
Vehicles. European Journal of Electrical Engineering and Computer
Science. 7(1), 38-46. https://doi.org/10.24018/ejece.2023.7.1.485
Vignesh, R., & Ashok, B. (2023). Intelligent energy management through
neuro-fuzzy based adaptive ECMS approach for an optimal battery
utilization in plugin parallel hybrid electric vehicle. Energy
Conversion and Management. 280(4), 116792-116793.
https://doi.org/10.1016/j.enconman.2023.116792
Vignesh, R., Ashok, B., Kumar, M.S., Szpica, D., Harikrishnan, A., & Josh,
H. (2023). Adaptive neuro fuzzy inference system-based energy
management controller for optimal battery charge sustaining in
biofuel powered non-plugin hybrid electric vehicle. Sustainable
Energy Technologies and Assessments. 59(10), 103379-103380.
https://doi.org/10.1016/j.seta.2023.103379
Wang, Y., Wu, Y., Tang, Y., Li, Q., & He, H., (2023). Cooperative energy
management and eco-driving of plug-in hybrid electric vehicle via
multi-agent reinforcement learning. Applied Energy, 332(10),
120563-120564. https://doi.org/10.1016/j.apenergy.2022.120563
Yang, C., Du, X., Wang, W., Yuan, L., & Yang, L. (2024). Variable
optimization domain-based cooperative energy management
strategy for connected plug-in hybrid electric vehicles. Energy,
290(4), 130206-130207.
https://doi.org/10.1016/j.energy.2023.130206
Yao, Z., Yoon, H.S., & Hong, Y.K. (2023). Control of hybrid electric
vehicle powertrain using offline-online hybrid reinforcement
learning. Energies, 16(2), 652-653.
https://doi.org/10.3390/en16020652
Younes, D., Karim, N., & Boudiaf, M. (2023). Energy management based
hybrid fuel cell/battery for electric vehicle using type 2 fuzzy logic
controller. International Journal of Advanced Studies in Computer
Science and Engineering, 12(1), 18-33.
https://doi.org/10.1109/ICEIT48248.2020.9113162
Zhang, Q., Tian, S., & Lin, X. (2023). Recent advances and applications
of ai-based mathematical modeling in predictive control of hybrid
Y.Ge
Int. J. Renew. Energy Dev 2024, 13(6), 1104-11 14
| 1114
ISSN: 2252-4940/© 2024. The Author(s). Published by CBIORE
electric vehicle energy management in China. Electronics, 12(2),
445-446. https://doi.org/10.3390/electronics12020445
.
© 2024. The Author(s). This article is an open access article distributed under the terms and conditions of the Creative Commons
Attribution-ShareAlike 4.0 (CC BY-SA) International License (http://creativecommons.org/licenses/by-sa/4.0/)
