Developing Equivalent Consumption Minimization Strategy for Advanced Hybrid System-II Electric Vehicles

: Compared with conventional vehicles, hybrid electric vehicles (HEVs) have the advantage of high-energy conversion e ﬃ ciency, which can have better fuel economy and lower emissions. The main issue of HEVs is how to develop an energy management strategy to achieve signiﬁcantly better fuel e ﬃ ciency. In this research, the Equivalent Consumption Minimization Strategy (ECMS) was applied to optimize the performance of fuel consumption in the Advanced Hybrid System-II (AHS-II). Based on FTP-75 Test Procedure deﬁned by the U.S. Environmental Protection Agency (EPA), a backward simulation module was established. The baseline simulation module with the rule-based control strategy was validated with the original fuel consumption data. Then, the module with ECMS followed the same control rules of engine on / o ﬀ and mode selection, and the fuel consumption of ECMS was compared with the simulation results of the baseline model. The fuel economy improvements of ECMS in urban, highway driving pattern, and composite fuel economy were up to 8.5%, 7.7%, and 8.1%, respectively. The simulation results showed that the di ﬀ erence of motors’ working e ﬃ ciency was only 1.2% between ECMS and baseline rule-based control strategies. The main reason of fuel consumption improvement was the engine operation chosen by ECMS, which provided better power distribution.


Introduction
Climate change and the sustainable development of energy are the most serious international issues in the 21st century. The hybrid electric vehicle (HEV) is one of the key technologies for vehicle energy saving. Combining with the internal combustion engine (ICE) and high efficiency electric motor, hybrid electric vehicles have better fuel economy than traditional vehicles. The HEVs are also more achievable than electric vehicles (EVs) under current limitations of battery.
With ICE and electric motors, the operating modes to control engine and motors are necessary designs for the hybrid power system, and the switching between the modes would change its power flow and the operation of the components. A planetary gear set (PGS) is often applied to the configuration of HEV. The most representative design is Toyota Prius released in 1997. In 2005, the Advanced Hybrid System-II (AHS-II), also known as the two-mode hybrid system, was developed by General Motors. AHS-II offers an additional set of electric-continue-variable-transmission (eCVT) mode of operation, and significantly reduce the energy loss in high speed [1]. Arata et al. [2] analyzed two different power-split hybrid-electric vehicle (HEV) powertrains using backward-looking simulations, and compared the Toyota Hybrid System II (THS-II) and the General Motors Allison Hybrid System II (GM AHSII).
Fuel economy and lower pollution emissions are critical issues. Vehicle manufactures are investing in energy management strategy (EMS) to ensure that the system components operate in their safe

Introduction of Hybrid Powertrain System
The AHS-II two-mode hybrid powertrain consists of a planetary gear, a compound planetary gear, four clutches, an internal combustion engine, and two motor/generators, MG1 and MG2, as shown in Figure 1. The architecture can produce effects which are similar to the continuously variable transmission (CVT). Therefore, it is also known as electrical continuously variable transmission (eCVT). The symbols R, S, and C represent the ring gear, the sun gear, and the carrier, respectively, and the subscripts 1 and 2 represent the compound planetary gear set and the simple planetary gear set. When the vehicle travels on different road conditions, the powertrain system can operate between two eCVT modes and four fixed-gear modes.

HEV Simulation Model
The backward calculation dynamic model of the HEV system was implemented using Matlab/Simulink, as shown in Figure 2. The US FTP-75 (EPA Federal Test Procedure) urban and highway drive cycles were applied as the road conditions for simulation, as shown in Figures 3 and 4. According to the known driving speed, the vehicle dynamic model calculated the required vehicle acceleration and driving torque, and through the energy management control module, the operating mode of the system and the output powers of engine and motor/generators, MG1 and MG2, were determined. The speed and torque of the two motor/generators were determined by the gear ratio of the transmission module. According to the speed and torque of the motor/generators, the battery module calculated the state of charge (SOC) of the battery pack. Finally, the fuel consumption was accumulated by ICE module through an engine two-dimensional lookup table.

HEV Simulation Model
The backward calculation dynamic model of the HEV system was implemented using Matlab/Simulink, as shown in Figure 2. The US FTP-75 (EPA Federal Test Procedure) urban and highway drive cycles were applied as the road conditions for simulation, as shown in Figures 3 and 4. According to the known driving speed, the vehicle dynamic model calculated the required vehicle acceleration and driving torque, and through the energy management control module, the operating mode of the system and the output powers of engine and motor/generators, MG1 and MG2, were determined. The speed and torque of the two motor/generators were determined by the gear ratio of the transmission module. According to the speed and torque of the motor/generators, the battery module calculated the state of charge (SOC) of the battery pack. Finally, the fuel consumption was accumulated by ICE module through an engine two-dimensional lookup table.

HEV Simulation Model
The backward calculation dynamic model of the HEV system was implemented using Matlab/Simulink, as shown in Figure 2. The US FTP-75 (EPA Federal Test Procedure) urban and highway drive cycles were applied as the road conditions for simulation, as shown in Figures 3 and 4. According to the known driving speed, the vehicle dynamic model calculated the required vehicle acceleration and driving torque, and through the energy management control module, the operating mode of the system and the output powers of engine and motor/generators, MG1 and MG2, were determined. The speed and torque of the two motor/generators were determined by the gear ratio of the transmission module. According to the speed and torque of the motor/generators, the battery module calculated the state of charge (SOC) of the battery pack. Finally, the fuel consumption was accumulated by ICE module through an engine two-dimensional lookup table.

HEV Simulation Model
The backward calculation dynamic model of the HEV system was implemented using Matlab/Simulink, as shown in Figure 2. The US FTP-75 (EPA Federal Test Procedure) urban and highway drive cycles were applied as the road conditions for simulation, as shown in Figures 3 and 4. According to the known driving speed, the vehicle dynamic model calculated the required vehicle acceleration and driving torque, and through the energy management control module, the operating mode of the system and the output powers of engine and motor/generators, MG1 and MG2, were determined. The speed and torque of the two motor/generators were determined by the gear ratio of the transmission module. According to the speed and torque of the motor/generators, the battery module calculated the state of charge (SOC) of the battery pack. Finally, the fuel consumption was accumulated by ICE module through an engine two-dimensional lookup table.

Vehicle Dynamic Module
According to the driving cycle, the vehicle speed was received, and the road load of vehicle model, including rolling resistance, aerodynamic drag, grade resistance, was calculated by Equations (1)- (4). Further, the AHS-II output shaft required torque was obtained. The vehicle parameters are shown in Table 1.
where FLoad, Fr, Fw, and Fg are vehicle road load, rolling resistance, aerodynamic drag, and grade resistance, respectively, and fr, M, g, α, ρ, Af, CD, and V are rolling resistance coefficient, vehicle mass, gravity, road slope, air density, vehicle front area, aerodynamic drag coefficient, and vehicle speed, respectively.

Controller Module
In this study, two sets of controllers were established, the rule-based control module for baseline HEV model and the Equivalent Consumption Minimization Strategy (ECMS) module for the optimized HEV model. These control modules were combined with the engine switch strategy to

Vehicle Dynamic Module
According to the driving cycle, the vehicle speed was received, and the road load of vehicle model, including rolling resistance, aerodynamic drag, grade resistance, was calculated by Equations (1)-(4). Further, the AHS-II output shaft required torque was obtained. The vehicle parameters are shown in Table 1.
where F Load , F r , F w , and F g are vehicle road load, rolling resistance, aerodynamic drag, and grade resistance, respectively, and f r , M, g, α, ρ, A f , C D , and V are rolling resistance coefficient, vehicle mass, gravity, road slope, air density, vehicle front area, aerodynamic drag coefficient, and vehicle speed, respectively.

Controller Module
In this study, two sets of controllers were established, the rule-based control module for baseline HEV model and the Equivalent Consumption Minimization Strategy (ECMS) module for the optimized HEV model. These control modules were combined with the engine switch strategy to further determine the operating time of the ICE. The individual controllers were built with Matlab functions.
Based on the torque required for the vehicle driving and the battery SOC, the rule-based controller, heuristic method (if-then-else), determined the speed and torque of ICE. The ESCM with object function developed the working state of the engine. After determining the status of ICE, the mode switch control module would switch between different eCVT and fixed-gear modes.

Transmission Module
The transmission includes two planetary gear sets, two motor/generators, and four clutches. Based on the vehicle driving condition, the mode switch module would determine the mode of operation, mode 1 for first eCVT mode and mode 2 for second eCVT mode. For mode 1, the motor speeds and torques of MG1 and MG2 were simulated by Equations (5)- (8). Equations (9)-(12) were for mode 2 operation. where, ω e , ω MG1 , ω MG2 , ω out , T e , T MG1 , T MG2 , and T out are the rotational speeds and torques of the engine, two motors, and transmission output. R R1 , R R2 , R S1 and R S2 are the radii of ring gear 1 and 2 and of sun gear 1 and 2, respectively. i 1 , i 2 are the radius ratio of sun gear to ring gear for gear train 1 and 2, respectively.
In the simulation, the rotational inertia of engine, I e ; inertia of ring gear 1 and 2, I R1 and I R2 ; inertia of carrier 1 and 2, I C1 , and I C2 ; inertia of motor/generator 1 and 2, I MG1 and I MG2 ; and inertia of sun gear 1 and 2, I S1 and I S2 ; are all considered. In mode 1 case, the general force-acceleration matrix can be written as shown in Equation (15). Similarly, Equation (16) is the case of mode 2.
Energies 2020, 13, 2033 where F 1 , F 2 , and F tire are the forces acting on the sun gear, ring gear, and tire, respectively. K f is the final axle ratio, and r tire is the radius of tire. I wheel is the total rotational inertia of the wheels [18].

Internal Combustion Engine Module
The ICE module of this study was represented by a lookup table. Figure 5 shows the three-dimensional ICE fuel consumption rate. Through the controller module to determine the engine running state, the corresponding engine speed and torque could determine engine fuel consumption rate.
F1, F2, and Ftire are the forces acting on the sun gear, ring gear, and tire, respectively. Kf is the final axle ratio, and rtire is the radius of tire. Iwheel is the total rotational inertia of the wheels [18].

Internal Combustion Engine Module
The ICE module of this study was represented by a lookup table. Figure 5 shows the threedimensional ICE fuel consumption rate. Through the controller module to determine the engine running state, the corresponding engine speed and torque could determine engine fuel consumption rate.

Motor/Generator Module
In AHS-II powertrain, there are two electric motor/generators, MG1 and MG2, which have same output power. The motor/generators are 60kW permanent magnet AC motors. In this study, MG1 and MG2 had same specifications. The motor/generators efficiency is a function of speed and output torque, as shown in Figure 6. This module was modeled with a lookup table. The motor/generator power calculation is shown in Equation (18).

Motor/Generator Module
In AHS-II powertrain, there are two electric motor/generators, MG1 and MG2, which have same output power. The motor/generators are 60kW permanent magnet AC motors. In this study, MG1 and MG2 had same specifications. The motor/generators efficiency is a function of speed and output torque, as shown in Figure 6. This module was modeled with a lookup table. The motor/generator power calculation is shown in Equation (18).
where P MG , ω MG , and T MG are motor/generators power, speed, and torque, respectively. If the speed and torque are in the same sign, the motor/generator works as a motor. If the speed and torque are in different sign, the motor/generator works as a generator, which transforms the mechanical energy into electricity and stores in the battery pack. η MG is the efficiency of the motor/generator, and K is the power flow of the motor/generator. K = 1 is motoring, and K = −1 is generating. where PMG, ωMG, and TMG are motor/generators power, speed, and torque, respectively. If the speed and torque are in the same sign, the motor/generator works as a motor. If the speed and torque are in different sign, the motor/generator works as a generator, which transforms the mechanical energy into electricity and stores in the battery pack. ηMG is the efficiency of the motor/generator, and K is the power flow of the motor/generator. K = 1 is motoring, and K = −1 is generating. Figure 6. The efficiency of motor/generator.

Battery Module
This study used a battery equivalent circuit, as shown in Figure 7, to establish the battery module [13]. The model provides the information of the open circuit voltage, output voltage, battery current, required power of battery, battery SOC, SOC changing rate, battery internal resistance, and battery capacity. The battery was mainly to support the power required for the motor/generator in order to keep the system in high fuel efficiency. The proposed model was adequate for the fuel consumption optimization.

Battery Module
This study used a battery equivalent circuit, as shown in Figure 7, to establish the battery module [13]. The model provides the information of the open circuit voltage, output voltage, battery current, required power of battery, battery SOC, SOC changing rate, battery internal resistance, and battery capacity. The battery was mainly to support the power required for the motor/generator in order to keep the system in high fuel efficiency. The proposed model was adequate for the fuel consumption optimization.
where PMG, ωMG, and TMG are motor/generators power, speed, and torque, respectively. If the speed and torque are in the same sign, the motor/generator works as a motor. If the speed and torque are in different sign, the motor/generator works as a generator, which transforms the mechanical energy into electricity and stores in the battery pack. ηMG is the efficiency of the motor/generator, and K is the power flow of the motor/generator. K = 1 is motoring, and K = −1 is generating.

Battery Module
This study used a battery equivalent circuit, as shown in Figure 7, to establish the battery module [13]. The model provides the information of the open circuit voltage, output voltage, battery current, required power of battery, battery SOC, SOC changing rate, battery internal resistance, and battery capacity. The battery was mainly to support the power required for the motor/generator in order to keep the system in high fuel efficiency. The proposed model was adequate for the fuel consumption optimization.   In Figure 7, V oc is the open circuit voltage, R eq is the internal equivalent resistance, and V L is the output voltage. The power output of battery can be represented in terms of the electric current, I batt , as shown in Equation (19). The total output of battery, V oc I batt , includes the power required for the system, P em_batt , and the power consumed by the internal resistance of the battery, R eq I 2 batt . The power required can also represented in terms of motors' power, as shown in Equation (20).
where η MG1 and η MG2 are the efficiency of the motor/generator 1 and 2, η con is the motor controller efficiency. The battery SOC can be calculated by accumulating the charged and discharged current.
The relationship between battery SOC changing rate, battery capacity Q max and current I batt is as follows: Energies 2020, 13, 2033 8 of 19 From Equation (19), the battery current can be derived as follows: The open circuit voltage and internal equivalent resistance are function of SOC. From Equations (21) and (22), the SOC rate can be obtained as follows: The internal resistance of battery was based on the curve shown in Figure 8. A portion of the battery power output was provided for the driving system, and the other was consumed by the internal resistance. The efficiency of battery can be calculated, as shown in Equation (24).
where ηMG1 and ηMG2 are the efficiency of the motor/generator 1 and 2, ηcon is the motor controller efficiency. The battery SOC can be calculated by accumulating the charged and discharged current. The relationship between battery SOC changing rate, battery capacity Qmax and current Ibatt is as follows: From Equation (19), the battery current can be derived as follows: eq batt em eq oc oc batt The open circuit voltage and internal equivalent resistance are function of SOC. From Equations (21) and (22), the SOC rate can be obtained as follows: The internal resistance of battery was based on the curve shown in Figure 8. A portion of the battery power output was provided for the driving system, and the other was consumed by the internal resistance. The efficiency of battery can be calculated, as shown in Equation (24). In this study, the SOC of battery was limited in the range between 0.4 and 0.6 since the battery had relatively less energy loss due to the battery internal resistance while considering for both charging and discharging states.

Optimization
The objective of the optimization problem was to minimize the fuel consumption and satisfy the following requirements for the HEV system: (1) To meet the demand of vehicle driving condition, and (2) to be constrained within the operation limits of the system components, as shown in In this study, the SOC of battery was limited in the range between 0.4 and 0.6 since the battery had relatively less energy loss due to the battery internal resistance while considering for both charging and discharging states.

Optimization
The objective of the optimization problem was to minimize the fuel consumption and satisfy the following requirements for the HEV system: (1) To meet the demand of vehicle driving condition, and (2) to be constrained within the operation limits of the system components, as shown in Equations (25)-(30). The goal of the optimization, the cost function J, is expressed numerically in finite time, as shown in Equation (25). For the hybrid powertrain system with charge-sustaining control, the initial battery SOC and the final state should remain the same. In other words, the power loss of the system must be compensated by the engine. The power required for the vehicle is provided through engine and motors, as shown in Equation (26). Equation (27)-(30) define the SOC controlled limits, battery power output limits, engine power output limits, and motor power output limits, respectively. The optimization problem is defined as the following: Objective: Subject to P req (t) = P e (t) + P em (t), where t, J, m fc (t), P batt , P e , P em , and P req are time, cost function, engine fuel rate, battery power, electric motor power, engine power, and vehicle power required, respectively. In this study, the SOC was limited between 0.4 and 0.6. Battery power and electric motor power were constrained between −60 kW and 60 kW. The engine power was between 0 kW and 157 kW.

Rule-Based Control Strategy
According to the understanding of the system architecture and the efficiency of each element, the output energy of each driving element is defined based on the different road conditions. The basic principle is to meet the driving force required during vehicle travelling, while the control rule should keep the engine and motor/generators in the high operating efficiency range as long as possible to achieve the best fuel consumption and the lowest emissions. This study applied a rule-based controller for the baseline HEV model, and the heuristic was applied, as shown in Figure 9. The fuel economy of this controller would be compared with the manufacture data to verify the accuracy the HEV model. The rule-based strategy is listed in Table 2. Based on different SOC and required torque output, engine operation conditions are provided.
through engine and motors, as shown in Equation (26). Equation (27)-(30) define the SOC controlled limits, battery power output limits, engine power output limits, and motor power output limits, respectively. The optimization problem is defined as the following: Objective: where t, J, mfc(t), Pbatt, Pe, Pem, and Preq are time, cost function, engine fuel rate, battery power, electric motor power, engine power, and vehicle power required, respectively. In this study, the SOC was limited between 0.4 and 0.6. Battery power and electric motor power were constrained between −60 kW and 60 kW. The engine power was between 0 kW and 157 kW.

Rule-Based Control Strategy
According to the understanding of the system architecture and the efficiency of each element, the output energy of each driving element is defined based on the different road conditions. The basic principle is to meet the driving force required during vehicle travelling, while the control rule should keep the engine and motor/generators in the high operating efficiency range as long as possible to achieve the best fuel consumption and the lowest emissions. This study applied a rule-based controller for the baseline HEV model, and the heuristic was applied, as shown in Figure 9. The fuel economy of this controller would be compared with the manufacture data to verify the accuracy the HEV model. The rule-based strategy is listed in Table 2. Based on different SOC and required torque output, engine operation conditions are provided. Figure 9. Heuristic controller. Figure 9. Heuristic controller. energy loss due to the internal resistance while the SOC is within 0.4 and 0.6, the design strategy uses this interval as SOC limits. In ECMS, the mode of operation should be determined first. Since the mode 2 is applied to the higher speed, if the SOC does not exceed 0.6, the engine will continue to operate to ensure that the battery system has enough power.
For mode 1 operation, when the vehicle speed is less than 20 km/h, the operating efficiency of engine will be in poor condition. If the SOC is not lower than 0.45, the engine will be shut down and vehicle is driven by electric motor to enhance fuel consumption performance. When the vehicle speed is between 20 km/h and 40 km/h, and the SOC is greater than or equal to 0.55, the engine will be shut down. When the vehicle speed is greater than or equal to 40 km/h, and the SOC is greater than or equal to 0.6, the engine will be shut-down. If SOC is greater than 0.6 and engine remains off, the SOC is monitored and checked until SOC is less than 0.55 and the process is reset to the starting block. The control flowchart of the engine switch is shown in Figure 10. This control logic was applied on the energy management strategies of both HEV models in this study.

Engine Switch Control Strategy
In this study, the engine switch was designed to avoid engine operating in high fuel consumption regions and to avoid the overcharging and discharging of the battery pack. Since the battery has less energy loss due to the internal resistance while the SOC is within 0.4 and 0.6, the design strategy uses this interval as SOC limits. In ECMS, the mode of operation should be determined first. Since the mode 2 is applied to the higher speed, if the SOC does not exceed 0.6, the engine will continue to operate to ensure that the battery system has enough power.
For mode 1 operation, when the vehicle speed is less than 20 km/h, the operating efficiency of engine will be in poor condition. If the SOC is not lower than 0.45, the engine will be shut down and vehicle is driven by electric motor to enhance fuel consumption performance. When the vehicle speed is between 20 km/h and 40 km/h, and the SOC is greater than or equal to 0.55, the engine will be shut down. When the vehicle speed is greater than or equal to 40 km/h, and the SOC is greater than or equal to 0.6, the engine will be shut-down. If SOC is greater than 0.6 and engine remains off, the SOC is monitored and checked until SOC is less than 0.55 and the process is reset to the starting block. The control flowchart of the engine switch is shown in Figure 10. This control logic was applied on the energy management strategies of both HEV models in this study.

HEV Baseline Model with Rule-Based Control Strategy
The rule-based control was applied on the baseline HEV model. With the urban and highway driving simulation, fuel economies were 47 MPG (miles per gallon) and 39 MPG, and composite fuel economy was 43 MPG. The simulation results and the manufacture official data are shown in Table  3. The differences of urban, highway, and composite fuel economy were −2.1%, 5.4%, and 1.6%,

HEV Baseline Model with Rule-Based Control Strategy
The rule-based control was applied on the baseline HEV model. With the urban and highway driving simulation, fuel economies were 47 MPG (miles per gallon) and 39 MPG, and composite fuel economy was 43 MPG. The simulation results and the manufacture official data are shown in Table 3. The differences of urban, highway, and composite fuel economy were −2.1%, 5.4%, and 1.6%, respectively, which were acceptable. The simulation model can be used to represent the original vehicle. With this rule-based control model as the baseline, the average efficiency of MG1 and MG2 were monitored as well. The urban and highway efficiency values of MG1 were 0.83 and 0.85, respectively, and those of MG2 were 0.85 and 0.84, as shown in Table 4.

HEV Optimization Model with ECMS
The ECMS was applied as optimization strategy for fuel economy simulation. The result is shown in Table 5. With ECMS, the fuel economy of urban and highways driving cycles had improvements around 8%, while the efficiencies of MG1 and MG2 were tracked as well, as shown in Table 6. The average of urban and highway efficiency values of MG1 were 0.83 and 0.85, and those of MG2 were 0.85 and 0.83, respectively. The difference between the baseline and ECMS was the MG2 highway efficiency, 84% vs. 83%, which was around 1.2% different. Under two different control strategies, the motor/generator efficiencies were very much the same. That indicates the improvement of fuel economy using ECMS was mainly due to the selection of the engine operating points.  Figure 11 to Figure 12 show the engine operating points of baseline and ECMS models. With ECMS optimization, the distribution of the engine operating points in urban and highway driving cycles was significantly smaller than that of the rule-based control strategy. With less engine power during the driving cycles, the ECMS model had less fuel consumption, as shown in Figures 13 and 14.            In highway driving cycle, as shown in Figures 17 and 18, the engine load with the rule-based control strategy had a significant operating time ratio in the less working efficiency range, accounting for about 22% of the engine operating time, while ECMS did not operate at all in the less efficiency range. With respect to overall operating time, ECMS optimization had a longer running time when operating in the better efficiency range, so HEV model with ECMS optimization could obtain better fuel economy.  In highway driving cycle, as shown in Figures 17 and 18, the engine load with the rule-based control strategy had a significant operating time ratio in the less working efficiency range, accounting for about 22% of the engine operating time, while ECMS did not operate at all in the less efficiency range. With respect to overall operating time, ECMS optimization had a longer running time when operating in the better efficiency range, so HEV model with ECMS optimization could obtain better fuel economy. In highway driving cycle, as shown in Figures 17 and 18, the engine load with the rule-based control strategy had a significant operating time ratio in the less working efficiency range, accounting for about 22% of the engine operating time, while ECMS did not operate at all in the less efficiency range. With respect to overall operating time, ECMS optimization had a longer running time when operating in the better efficiency range, so HEV model with ECMS optimization could obtain better fuel economy.

Conclusions
Energy management strategy is an important topic for hybrid electric vehicles. In order to effectively improve the performance of fuel consumption, ECMS was selected for instant optimization strategy to optimize the fuel consumption.
The AHS-II two-mode hybrid system was modeled, and a rule-based control was used as baseline model and validated with the official fuel economy data. Then, the ECMS was applied for fuel economy optimization. From the simulation of baseline model and ECMS optimization, the conclusions were drawn as follows: 1. For baseline model with the rule-based control strategy, the fuel economy of the urban, highway, and composite driving cycles were 47 MPG, 39 MPG, and 43 MPG. The official reported fuel economy data of AHS-II are 48 MPG, 37 MPG, and 42 MPG for the urban, highway, and composite driving cycle, respectively. The maximum difference of fuel economy between simulation and vehicle official data is 5.4%, which indicates that the results of simulation model has good correlation with those of the vehicle, and can be used to represent the original vehicle. 2. The fuel economies with ECMS optimization were 51 MPG, 42 MPG, and 46.5 MPG for urban, highway, and composite driving cycles, respectively. Comparing ECMS optimization with the rule-based control strategy, the improvements made were 8.5%, 7.7%, and 8.1%, separately. The proposed ECMS optimization strategy provided a better fuel economy performance. 3. The efficiencies of the motor/generator 1 and 2 in the rule-based model were 0.83 and 0.85 for city, and 0.85 and 0.84 for the highway driving cycles. Those in ECMS were 0.83 and 0.85 for city, and 0.85 and 0.83 for highway, individually. The biggest difference between the baseline and ECMS was only 1.2%. The improvement of fuel economy was mainly due to the selection of the engine operating points which lead to a better fuel performance. ECMS could effectively implement the best engine power distribution to achieve better fuel consumption. 4. In urban driving cycle, there was 42% of time that engine was operated in the less efficiency region for the rule-based control strategy, while there was 27% with ECMS. 5. In highway driving cycle, there was 22% of the time that engine was operated in the less efficiency region for the rule-based control strategy, while there were none for ECMS. Overall, ECMS optimization had engine operated in better efficiency range and provided better fuel economy.

Conclusions
Energy management strategy is an important topic for hybrid electric vehicles. In order to effectively improve the performance of fuel consumption, ECMS was selected for instant optimization strategy to optimize the fuel consumption.
The AHS-II two-mode hybrid system was modeled, and a rule-based control was used as baseline model and validated with the official fuel economy data. Then, the ECMS was applied for fuel economy optimization. From the simulation of baseline model and ECMS optimization, the conclusions were drawn as follows: 1.
For baseline model with the rule-based control strategy, the fuel economy of the urban, highway, and composite driving cycles were 47 MPG, 39 MPG, and 43 MPG. The official reported fuel economy data of AHS-II are 48 MPG, 37 MPG, and 42 MPG for the urban, highway, and composite driving cycle, respectively. The maximum difference of fuel economy between simulation and vehicle official data is 5.4%, which indicates that the results of simulation model has good correlation with those of the vehicle, and can be used to represent the original vehicle.

2.
The fuel economies with ECMS optimization were 51 MPG, 42 MPG, and 46.5 MPG for urban, highway, and composite driving cycles, respectively. Comparing ECMS optimization with the rule-based control strategy, the improvements made were 8.5%, 7.7%, and 8.1%, separately. The proposed ECMS optimization strategy provided a better fuel economy performance.

3.
The efficiencies of the motor/generator 1 and 2 in the rule-based model were 0.83 and 0.85 for city, and 0.85 and 0.84 for the highway driving cycles. Those in ECMS were 0.83 and 0.85 for city, and 0.85 and 0.83 for highway, individually. The biggest difference between the baseline and ECMS was only 1.2%. The improvement of fuel economy was mainly due to the selection of the engine operating points which lead to a better fuel performance. ECMS could effectively implement the best engine power distribution to achieve better fuel consumption. 4.
In urban driving cycle, there was 42% of time that engine was operated in the less efficiency region for the rule-based control strategy, while there was 27% with ECMS.

5.
In highway driving cycle, there was 22% of the time that engine was operated in the less efficiency region for the rule-based control strategy, while there were none for ECMS. Overall, ECMS optimization had engine operated in better efficiency range and provided better fuel economy.
Using vehicle dynamic simulation can help to quickly carry out the preliminary evaluation of various energy management strategies, search for the appropriate optimization of energy management strategies, and reduce R&D process time and cost. The proposed equivalent consumption minimization strategy can be utilized to optimize the performance of fuel consumption.
Funding: This research received no external funding.

Conflicts of Interest:
The author declares no conflict of interest. summation of instant fuel consumption PHEV plug-in hybrid electric vehicle P batt power output of battery P e output power of the engine P em output power of the electric motor