Printer Friendly

Analysis and Control of a Torque Blended Hybrid Electric Powertrain with a Multi-Mode LTC-SI Engine.

INTRODUCTION

Two-thirds of the petroleum consumption around the world is consumed in the transportation sector and half of that goes to passenger cars and light duty trucks [1]. Petroleum comprised 92% of U.S. transportation energy use in 2014 and prevalent consumption of petroleum-based fuels leads to high Greenhouse Gas (GHG) emission. For instance, in 2012, the European transportation sector was responsible for 19.7% of total GHG emissions and the aim is to cut it by 60% in 2050 [2]. In this context, automakers must reduce GHG emissions by introducing advanced fuel-efficient technologies and also by using alternative fuels. Low Temperature Combustion (LTC) engines encompass a family of Internal Combustion Engine (ICE) technologies, such as Homogeneous Charge Compression Ignition (HCCI), and Reactivity Controlled Compression Ignition (RCCI) engines, which have low engine-out NOx and low soot emissions, and offer peak net indicated thermal efficiency as high as 53% [3]. Moreover, improvement in ICE fuel efficiency has the largest potential in improving Hybrid Electric Vehicles (HEVs) fuel economy and reducing the GHG emission, compared to enhancement in other HEV component efficiency [4]. Thus, integrating fuelefficient LTC engines in HEVs has the potential to improve vehicle fuel economy and decrease GHG emission.

However, utilizing the LTC engines in vehicles faces two major challenges: (i) limited engine operating range, and (ii) combustion control complexity particularly during transient engine operation. To tackle these two challenges, this paper investigates integration of an LTC engine with a parallel Hybrid Electric Vehicle (HEV) powertrain by taking advantage of torque assist from the electric motor. This allows the engine to be downsized and the engine operating points shift to the low and mid power level and operate in a narrow LTC operating range. In addition, this also significantly reduces the engine transient operation. Moreover, the parallel HEV is likely to be the dominant strong hybrid technology based on projected cost and effectiveness versus the currently-dominant power-split design [5]. There are mainly three types of engines used in HEV architecture. These three types include: Compression Ignition (CI) engines, SI engines, and LTC engines. Figure 1 divides prior HEV studies for the major ICE types.

In the first category, a CI engine has been used in different HEV architectures [6, 7, 8, 9, 10, 11, 12, 13, The CI engines have been mostly used in Sport Utility Vehicles (SUVs) and trucks [8, 10, 12, 12, 15]. In another category, SI engines have been used in HEV and EREV (Extended Range Electric Vehicle) configurations [16, 17, 18, 19, 20, 21]. Recently, Atkinson SI engines have become more popular in the market (e.g., Ford C-Max, Toyota Prius, Lexus RX 450h, and Honda Accord PHEV). In the study in reference [22], Toyota R&D group redesigned an engine based on the Atkinson cycle to improve the engine fuel efficiency. In addition, in [23], a Honda Accord PHEV was redesigned based on the Atkinson cycle and 10% lower fuel consumption was reported compared to the SI engine.

Few studies are found in the literature that investigated integrating LTC engines with HEVs. Among different types of LTC engines, HCCI was the first type that was explored in hybrid electric powertrains. In the first study at U.S. Argonne National Lab, the effects of using a dual-mode SI-HCCI engine in different vehicle electrification levels were analyzed [24]. Their simulation results predicted a 6% to 15% fuel consumption reduction, depending on powertrain configurations and driving cycles. In another study [25], fuel economy improvement of an HCCI engine versus an SI engine for both mild and medium parallel HEVs were investigated. The authors reported a 17% to 35% fuel economy improvement by using a dual-mode SI-HCCI in comparison to a conventional SI non-hybrid powertrain. In both studies [24, 25], a dual-mode (SI-HCCI) engine is used with a rule-based controller. In [26], we carried out the first study by utilizing a pure HCCI mode engine in a series hybrid powertrain and it was found 12.6% improvement in fuel economy in comparison with a series HEV running with an SI engine. In our next study [27], the impact of driving cycles, number of the engine operating points, and engine startup fuel penalty on both series HEV and EREV HCCI-based powertrains are investigated. The results show that fuel economy can increase by 17.7% by including more operating points, while fuel economy improvement is low for the driving cycles with lower average power ([P.sub.ave] [less than or equal to] 4.5 kW). RCCI is the second type of LTC engines that have been recently studied in an HEV powertrain. In [28], researchers at the University of Wisconsin-Madison used an RCCI engine in a hybrid electric powertrain. The RCCI engine was integrated with a series hybrid powertrain based on a modified 2009 Saturn Vue. The authors tested the vehicle for the US Environmental Federal Agency (EPA) Highway Fuel Economy Test (HWFET) procedure to measure the highway fuel economy and emissions of the vehicle. Based on simulation results, the authors predicted a series-parallel RCCI-HEV configuration will lead to a 12% fuel economy improvement for the Chevrolet Volt. In reference [29] 3% fuel economy improvement of RCCI engines over diesel engines in a series HEV architecture are investigated.

Moreover, in [30] we carried out the first study on integrating a multi-mode LTC engine with a series HEV architecture. The multimode LTC engine was able to switch between the HCCI, RCCI and conventional SI modes by incorporating a fuel penalty. The results show a 9% to 13.1% fuel economy improvement, compared to an identical series HEV platform running with a single-mode SI engine.

Based upon the previous work, this paper presents the first study undertaken to investigate the fuel economy benefit of integrating a multi-mode LTC engine with a parallel HEV with an advanced optimal control strategy. The LTC modes in this work include both HCCI and RCCI modes. The contribution from this study is threefold. First, it investigates the ultimate fuel consumption reduction of a multi-mode LTC-SI in parallel HEV configuration. Second, it investigates the effect of the hybridization level on the fuel saving over the single-mode SI in the parallel HEV. Third, it examines the trade-off between the number of engine modes on the powertrain fuel consumption.

This paper is organized as follows. In the first section, the HEV experimental setup is explained. Next, the optimal control problem in the parallel HEV architecture and Pontryagin's Minimum Principle (PMP) optimization techniques are applied. Then, the results of the parallel HEV for the single-mode and multi-mode engine operations are discussed. Finally, the last section summarizes all findings from this paper.

HYBRID ELECTRIC POWERTRAIN EXPERIMENTAL SETUP

This section introduces the experimental setup that is designed and built at Michigan Technological University for this study. The experimental setup enables testing LTC engine hybridization to investigate different aspects of the LTC-HEV powertrain. The setup is comprised of a fuel-flexible LTC engine and an electric powertrain, which are connected to a 465 hp double-ended AC dynamometer. Figure 2 shows the experimental setup in this study. The fuel-flexible LTC engine can be adapted to operate in different LTC modes including HCCI, RCCI and conventional SI modes. The electric powertrain is capable of realizing different levels of powertrain electrification and also has a proper control platform to implement HEV energy management strategies.

The next two subsections present the details of the LTC engine and electric powertrain experimental setups, respectively. In addition, powertrain testing conditions for developing the HEV component models are explained.

Engine Setup

The engine setup includes a GM 2.0 L Ecotec Gasoline Direct Injection Turbocharged SI engine. Table 1 lists the baseline SI engine specifications.

The baseline SI engine was converted to a multi-mode LTC-SI engine as shown in Figure 3. LTC modes include HCCI and RCCI operations. Major changes to the baseline SI engine include (i) design and programming new Engine Control Unit (ECU), (ii) adding port fuel injection systems, and (iii) capability to adjust intake charge temperature, pressure, and dilution level through utilizing intake air heater, supercharger, and Exhaust Gas Recirculation (EGR) rate modulation, respectively. In the following, the configuration of the multi-mode LTC-SI engine is explained.

The engine is controlled using dSPACE MicroAutoBox (MABx) DS1511 and RapidPro units. Models for control of injectors, fuel pump, spark plugs, throttle body, cam phasers, supercharger and EGR valve were developed in Matlab[R] Simulink. The intake and exhaust valves are actuated using cam phasors. All these models were integrated into a single engine management system, and these parameters were controlled and monitored online using the dSPACE ControlDesk[R]. Combustion analysis was performed using the ACAP system. The in-cylinder pressures were measured using PCB 115A04 piezo electric pressure transducers. The measured signal was amplified using a DSP 1104 CA charge amplifier and then sent to ACAP for post-processing. The encoder used for crank angle measurement had a resolution of 1 crank angle degree. ACAP was configured for the engine setup and the required combustion parameters were continuously evaluated for 100 cycles at each operating condition. National Instruments PXIe 1078 chassis system, which includes NI TB 4353, NI PXI 6225 and NI PXI 6722 modules were used for thermocouple measurement, pressure transducer measurement and dynamometer control, respectively.

Two fuel rails were installed at the interface of the intake manifold and cylinders for port fuel injection. A high-pressure fuel pump was used to supply fuel at 100-bar pressure to direct fuel injectors. In addition, a low-pressure external fuel pump was used to supply fuel at 3 bar pressure to the port fuel injectors. The injection system has the capability of supplying three fuels to the engine at the same time. The amount of fuel injected was controlled using dSPACE MABx. Temperatures (intake air, exhaust gas, coolant and oil gallery) and pressures (intake air, coolant and oil) were measured using K-type thermocouples with an accuracy of [+ or -]0.75% and piezo resistive pressure transducers with an accuracy of [+ or -]0.5%, respectively. Two air heaters between the supercharging station and the intake manifold were used to pre-heat the intake air to the desired temperature. The mass flow rate of intake air was measured using Merriam MDT500 air flow measurement system. A 465 hp General Electric AC dynamometer was used to control the speed and load of the engine.

Experiments were performed for three different combustion modes namely SI, HCCI and RCCI. Operating conditions such as intake air temperature ([T.sub.intake]), intake manifold pressure ([P.sub.intake]), Research Octane Number (RON) of fuel, engine speed (N) and equivalence ratio ([PHI]) were varied individually keeping the other parameters constant. The operating conditions used for each of the combustion modes are given in Table 2. The experiments were conducted for a range of engine speeds and a range of equivalence ratios between the knock and misfire limits. Thereby, the operating region and load limit for each combination of input parameters were determined. Using the data acquired from dSPACE[R], LabVIEW[R] and ACAP, the combustion and performance parameters were calculated using an in-house Matlab code. The Brake Specific Fuel Consumption (BSFC) maps were generated and the load limits for each of the combustion modes were determined. Figure 4 shows the BSFC maps for SI, HCCI and RCCI combustion modes with engine speed (RPM) on the x-axis and brake torque (N.m) on the y-axis.

Electric Powertrain Setup

The electric powertrain setup includes a 100 kW synchronous induction Remy e-motor, an RMS PM100DX inverter, a 5 kWh/65 kW lithium-ion LG Chem battery, and a mechanical coupling to integrate the motor to the dynamometer. The mechanical drivetrain including the e-motor mount, coupling, and shafts are designed and manufactured at Michigan Technological University. The high voltage battery during the operation is connected to the e-motor through a designed pre-charge circuit. The MABx is used as a supervisory controller to monitor sub-level controllers (i.e., battery, e-motor, etc.). The MABx communicates control commands on the CAN bus to the sub-level controllers. The LG Chem battery temperature is controlled through a fan and all the cooling systems are controlled by the supervisory controller. A fault-action module was developed in Matlab[R] to manage the setup during faults and extreme conditions.

The driver commanded torque is carried out through a desired torque-based control strategy embedded in the inverter. The driver's desired speed setpoint is controlled by the dyno controller. The inverter controls the e-motor to track the reference torque, which is determined by the operator through the dSPACE ControlDesk[R] interface. Furthermore, the setup is capable of implementing the regenerative braking test scenarios to capture the braking energy and charges the battery. The regenerative braking happens when the torque command is negative or the motor direction is reversed.

Figure 5 illustrates the LG Chem battery charging and discharging power limits versus SOC. The LG Chem battery SOC operating window is from 70 to 30 percent. As it is expected the battery charging power is smaller when the battery is fully charged. However, the battery has a higher discharging power limit at high SOC. This data will be used in the powertrain modeling for the design of the supervisory energy management strategy.

Figure 6 shows the e-motor efficiency map during different e-motor speeds and torques. The data is collected at the e-motor temperature of 70[degrees]C and the DC voltage of 320 V. Under these conditions, the e-motor efficiency ranges from 74 to 94 percent.

In the next section the designed optimal energy management strategy is presented. The optimal strategy is implemented for the parallel HEV model with the experimentally validated components using the collected data from the powertrain experimental setup.

DESIGN OF OPTIMAL CONTROL

The goal of the optimal control in this study is to minimize the fuel consumption ([??]) using tlie cost function defined by Eq. (1):

[mathematical expression not reproducible] (1)

where, [??] is the rate of the engine fuel consumption and T is the time length of a driving cycle. Equation (2) shows the constraints for the HEV optimization problem.

|[SOC.sub.f] - [SOC.sub.0]| [??] [epsilon] (2a)

[SOC.sub.min] [??] SOC(t) [??] [SOC.sub.max] (2b)

[P.sub.bat,min](SOC) [??] [P.sub.bat](t) [??] [P.sub.bat,max]{SOC) (2c)

[T.sub.eng,min]([[omega].sub.eng]) [??] [T.sub.eng](t, [[omega].sub.eng]) [??] [T.sub.eng,max]([[omega].sub.eng]) (2d)

[[omega].sub.eng,min] [??] [[omega].sub.eng](t) [??] [[omega].sub.eng,max] (2e)

[T.sub.motor,min]([[omega].sub.mot]) [??] [T.sub.motor](t) [??] [T.sub.motor,max]([[omega].sub.mot]) (2f)

[[omega].sub.motor,min] [??] [[omega].sub.motor](t) [??] [[omega].sub.motor,max] (2g)

The constraints in Eq. (2) are applied for the battery SOC operation window, battery power ([P.sub.bat]), engine torque ([T.sub.eng]), engine speed ([[omega].sub.eng]), e-motor torque ([T.sub.motor]), and e-motor speed ([[omega].sub.motor]). This optimal control problem is solved using optimal control techniques that are described in the next section.

Pontryagin's Minimum Principal (PMP)

The PMP method is based on a general case of the Euler-Lagrange equation and originates from the Calculus of Variation. This method yields the necessary - not sufficient - conditions of the global optimal solution. The optimal trajectories derived from PMP will be the global optimal solution of the HEV problem if the obtained optimal trajectory is a unique trajectory that meets the necessary and boundary conditions [31]. The necessary condition for the PMP global optimality is explained in [31] and will be briefly explained in Subsection B. In Subsection A, the simulation model of the target vehicle is described. Moreover, a technique is introduced to determine the Hamiltonian values from the model.

A. Parallel HEV Model

In the parallel HEV architecture in this study, the engine is coupled to the e-motor through a clutch. The output shaft is connected to the drivetrain where an automatic six speed transmission connects the output shaft to the wheels. This limits the engine operating points to discrete gear ratio options. Figure 7 shows the parallel HEV architecture and Table 3 lists the vehicle parameters along with the transmission ratios used in this study. The static models and maps of the vehicle components are obtained from the experimental setup as explained in the previous section.

In the parallel HEV in this work, there are four independent control variables. These variables include the engine torque, the battery power, the transmission gear ratio, and the clutch status for a given driver's power request and vehicle speed. Both the engine and e-motor speeds can be obtained for a given transmission gear ratio and vehicle speed. The e-motor and engine speeds are determined by Eq. (3), which are a function of the vehicle's speed and the transmission gear ratios as follows:

[mathematical expression not reproducible] (3)

where, [V.sub.veh] is the vehicle speed, [n.sub.t] is the transmission gear ratio, [n.sub.d] is the differential ratio, r is the wheel radius, and the [n.sub.c] is the engine/e-motor coupling ratio. In addition, for a given driver's power request the engine required power can be calculated by Eq. (4) to meet the driver's demand.

[P.sub.eng] = [P.sub.wheel] - ([P.sub.motor,mech]) (4)

Where, [P.sub.motor,mech] is the e-motor mechanical power. Since the battery is supplying the e-motor power, the [P.sub.motor,mech] can be obtained by knowing the battery power and the e-motor efficiency as follows

[mathematical expression not reproducible] (5)

Where, [P.sub.motor,e] is the e-motor input (electrical) power; [[eta].sub.motor] and [[eta].sub.gear] are the e-motor and the transmission efficiencies, respectively. k equals to -1 when the wheel power is positive and equals to 1 when the power at the wheel is negative. The wheel power is determined by the Longitudinal Vehicle Dynamic equations, as explained in the Appendix.

From Eq. (3) to (5), the engine power, motor power, engine speed, and motor speeds are specified by knowing the transmission gear ratio ([n.sub.t]) and the battery power at each time. The motor and engine torques are determined accordingly.

[mathematical expression not reproducible] (6)

The fuel consumption rate is determined from the engine BSFC map as a function of [T.sub.eng] and [[omega].sub.eng]:

[[??].sub.f] = F([T.sub.eng], [[omega].sub.eng]) (7)

Finally, the time derivative of SOC is obtained from the battery dynamic equation as follows

[mathematical expression not reproducible] (8)

where, [Q.sub.nom] is the battery nominal energy capacity. The equivalent open-circuit voltage (OCV) and internal resistance (R) are fiinctions of SOC. Thus, the SOC is a function of [P.sub.bat], and SOC. In conclusion, the [m.sub.f] in Eq. (7) and SOC in Eq. (8) depend on [P.sub.bat], and [n.sub.t], if the wheel power ([P.sub.wheel]) and vehicle speed are given. Moreover, the [P.sub.bat], and [n.sub.t] are selected as the control variables in the optimization problem. In this work, the battery and longitudinal vehicle dynamic (LVD) models are considered as dynamic models; however, the engine and e-motor are quasi-static map-based models which are extracted from experimental data.

B. Development of PMP-based torque management strategy

To apply the optimal control theory to the HEV powertrain, the Hamiltonian (H) is defined as follows:

[??]([P.sub.bat], SOC, t) = -[lambda](t).g([P.sub.bat] (t), SOC (t)) +[[??].sub.f] ([P.sub.bat], [n.sub.t], t) (9)

where, [lambda] is called "costate" in the PMP and g is the state equation, which encompasses the battery dynamics (Eq. 8). Thus, to calculate the Hamiltonian, first the SOC and g are determined using control variables, [P.sub.bat], and from Eq. (3) to (8).

Using the PMP optimization technique, the state equation and costate equation are obtained as

SOC = [[partial derivative]H/[partial derivative][lambda]] (10)

[??] = [[partial derivative]H/[partial derivative]SOC] (11)

From Eq. (9) and Eq. (11), if g is not a function of SOC then the costate can be considered constant, as explained below.

In refrence [30], it is shown that the battery OCV and battery resistance R do not vary significantly in the charge-sustaining over the battery SOC range i.e., from 0.3 to 0.7. In that case, the costate stays near the initial value since the [[partial derivative]g/[partial derivative]SOC] is negligible, compared to the costate for the whole driving cycle. Thus, the costate expression during the SOC usage window is simplified as

[??] = [lambda][[partial derivative]g/[partial derivative]SOC] =0, [right arrow] [lambda] = constant (12)

Obtaining a constant value for the costate reduces the PMP complexity. For optimality, the following condition should be considered to specify the optimal control variable [P.sub.bat] and [n.sub.t] at each time step

H ([P.sup.*.sub.bat], [n.sup.*.sub.t], [SOC.sup.*], [[lambda].sup.*], t) [??] H([P.sub.bat], [n.sub.t], [SOC.sup.*], [[lambda].sup.*], t) (13)

which means that the optimal control variables [P.sup.*.sub.bat] and [n.sup.*.sub.t] minimize the Hamiltonian function at the given time. The boundary condition of the final sate variable is

SOC([t.sub.0]) = SOC([t.sub.f]) (14)

The [P.sup.*.sub.bat], [n.sup.*.sub.t], and [[lambda].sub.*] that satisfy Eq. (13) and the boundary condition (14), determines the optimal [P.sub.bat] and [n.sub.t] trajectory. If a costate ([lambda]) exists that fulfills the condition (14), then the PMP provides the 'global optimal solution [31].

C. Extending the cost function for Multi-Mode operation

The cost function in Eq. (1) is revised by including the engine startup, LTC-SI mode-switching, and gear shifting fuel penalties for minimizing the number of engine starts, LTC-SI mode-switching, and gear-shifting. The revised cost function (J) is:

[mathematical expression not reproducible] (15)

[mathematical expression not reproducible] (16)

where, [GAMMA] determines the condition for adding the engine startup fuel penalty. [LAMBDA] is zero when the engine is enforced to operate in the single-mode regions and [LAMBDA] is 1 when the mode-switching is allowed. The engine startup fuel penalty [F.sub.p1] is measured 0.15 grams using data from the experimental setup previously explained. The [F.sub.p2] is considered 0.1 grams to smooth the transmission gear-shifting. [PSI] is 1 when [n.sub.t](t) [not equal to] [n.sub.t](t - 1) and [PSI] is zero when [n.sub.t](t) [not equal to] [n.sub.t](t - 1).

The[ m.sub.ij] is mode-switching fuel penalty to switch from the ith engine mode to the jth engine mode. i & j [member of] [1, 2, 3] since the engine can run in HCCI, RCCI, and SI modes. The [m.sub.ij] prevents frequent model-switching between different modes. It accounts for the fuel penalty associated with each mode switching and finds the global optimal solution. The [m.sub.ij] used in this study is an experimentally measured value and is taken from [32] for the mode-switching on a similar engine.

The constraints for the optimal control problem are revised based on the components in the powertrain experimental setup in this study:

|[SOC.sub.f] - [SOC.sub.0] [??] 0.01 (17a)

0.3 [??] SOC(t) [??] 0.7 (17b)

[P.sub.bat,min] [??] [P.sub.bat] (t) [??] [P.sub.bat,max] (17c)

[P.sub.eng,min] ([[omega].sub.eng]) [??] [P.sub.eng] (t, [[omega].sub.eng]) [??] [P.sub.eng,max] ([[omega].sub.eng]) (17d)

[[omega].sub.eng,min] [??] [[omega].sub.eng] (t) [??] [[omega].sub.eng,max] (17e)

0 [??] [P.sub.motor] (t) [??] 100 kW (17f)

0 [??] [[omega].sub.motor] (t) [??] 8000 RPM (17g)

where, a constraint of maximum one percent [DELTA]SOC variation is considered for the charge-sustaining mode. The next section presents the results for the single-mode engine and the multi-mode engine integrated with the parallel HEV powertrain. The PMP optimal control solution is used for the analysis in this work.

RESULTS

The PMP approach explained in the previous section is adopted to analyze and investigate the multi-mode LTC-SI engines in the parallel HEV configuration. The costate [lambda] in the PMP is selected such that the SOC is balanced at the end of the driving cycle, which means the initial and final SOCs are equal. Three different levels of hybridization are defined by the [[P.sub.bat]/[P.sub.eng]] ratio as listed in Table 4. This includes three levels based on the battery and engine power ratio. PHEV has the highest electrification level with peak 60 kW e-motor power and peak 410 V battery voltage. The SOC is also balanced for the PHEV electrification level assessment in this work. The battery and e-motor power limit in the full hybrid category is defined as 40 kW; this number reduces to 18 kW for the mild hybrid.

Figure 8 shows the single-mode SI engine BSFC map. The engine optimum operating points are shown for the Urban Dynamometer Driving Schedule (UDDS) and Highway Fuel Economy Test (HWFET) driving cycles. The engine operating points along with the hybridization levels are shown in Figure 8. In the UDDS driving cycle, the engine mostly operates in the low BSFC region for the PHEV, while the engine operating points shift to the low torque and high BSFC regions when the hybridization level decreases to mild hybrid category. A higher battery and e-motor power provides more flexibility for the hybrid powertrain to shift the engine operating points to the more efficient engine regions, while maintaining the battery SOC. However, a smaller e-motor in mild HEV limits the torque management strategy to shift the engine operating points to low engine BSFC region.

The engine speed in the parallel HEV is a function of the vehicle speed and the transmission gear ratio ([n.sub.t]). Thus, the torque management strategy tries to select [n.sub.t] and engine torque, which leads to the lowest BSFC value. But, the best BSFC point is not always feasible and the engine operating points cannot be placed in the low BSFC region (i.e., Figure 8-[a.sub.1]) due to the discrete choice of engine speeds. In HWFET driving cycle; however, the engine operates in the more efficient operating points compared to the UDDS driving cycle (i.e., compare Figure 8-[b.sub.2] with Figure 8-[a.sub.2]). The higher vehicle speed in the HWFET driving cycle helps the engine operating points to shift to the high speed region, even though the e-motor torque assist becomes limited in the mild hybrid.

Figure 9 shows the multi-mode LTC-SI engine BSFC map and the engine operating points over the UDDS and HWFET driving cycles. The engine operating points are illustrated for both PHEV and mild hybrid. Comparing Figures 8-[a.sub.2] and 9-[a.sub.2] shows that the high BSFC operating points in the single-mode engine are now running in the LTC modes (i.e., RCCI, HCCI) in the multi-mode engine. Increasing the running time of LTC modes reduces the overall fuel consumption of the vehicle. The time percentage of the vehicle different operating modes including EV, HCCI, RCCI, and SI are shown in Figure 10. As expected the EV mode decreases from PHEV to mild hybrid because the engine cannot charge the battery as aggressively as that in the PHEV mode.

In addition, by comparing Figure 10-[a.sub.1] and Figure 10-[b.sub.1] it can be seen that the engine running time in LTC modes increases rapidly by reducing the vehicle electrification level over the UDDS cycle. Moreover, comparing Figure 10-[a.sub.2] and Figure 10-[b.sub.2] shows that the LTC running time increases at a smaller rate by reducing the electrification level over the HWFET driving cycle. Among the LTC modes, the RCCI mode has the most engine running time due to the reasons that will be explained.

A detailed comparison of the engine operating modes is provided in Table 5. The table shows the fuel consumption for different hybridization levels along with the engine work and average engine efficiency for each case. The mild hybrid multi-mode LTC-SI case shows a 5% increase in fuel consumption, compared to the PHEV. Moreover, in the mild hybrid, the engine efficiency and engine work reduce by 4% and 1% respectively compared to the PHEV. Comparison between the multi-mode LTC-SI and single-mode SI shows that the single-mode SI engine has up to 8.8% higher fuel consumption, 9% lower engine efficiency, and 3% more engine work compared to the multi-mode LTC engine. However, the fuel consumption reduction in the multi-mode LTC-SI compared to the single-mode SI changes from 8.8% to 1.4% by increasing the hybridization level from mild HEV to PHEV.

In addition, Table 5 shows that the engine efficiency changes less in the multi-mode LTC-SI engine for different hybridization levels, compared to that in the single-mode SI engine. The relatively lower engine efficiency variation in the multi-mode case results from utilizing the engine in its LTC mode, which helps keep the average engine efficiency relatively high. This is the result of inherent characteristic of an LTC engine in which the high efficiency region is in the low to mid power area. With the high power battery in PHEV, the vehicle has flexibility for placing the engine operating points, independent of wheel power demand. As a result, the engine operating points are placed in high efficiency and high engine power regions to provide the propulsion power as well as charging the battery at its maximum capacity. However, in the multi-mode mild hybrid, the engine operating points cannot be placed in the engine high power region since the battery cannot be charged extensively due to its lower power capacity. Moreover, the engine ON time increases by reducing the electrification level. This also links to the less aggressive battery charge/discharge in the mild hybrid compared to the PHEV; thus, the vehicle electric mode decreases and the engine must run for a longer time to provide the demanded power.

In the multi-mode engine, however, the availability of high efficiency LTC operating points enforces the Hamiltonian function to choose the LTC operating modes, whereas in the single-mode the engine has to operate in lower efficiency regions. The advantage of keeping the engine efficiency close to PHEV is pronounced for low-power-demanding cycles such as the UDDS cycle. For highway cycles such as HWFET, since the peak wheel demand power is high at high vehicle velocities, in contrast to urban cycles that peak wheel demand usually happens at low vehicle speed. Thus, the HEV energy controller unit has the flexibility to operate the engine at high engine speeds with low BSFC values. The engine work difference between PHEV and mild hybrid in Table 5 and Table 6 are rooted in missing some regenerative power in the mild hybrid (due to battery power limits). Because of more aggressive battery charging/discharging, in general, the PHEV has more battery loss compared to mild hybrid in the both UDDS and HWFET cycles. Thus, the multi-mode LTC engine benefits more in the mild hybrid compared to the full hybrid and PHEV.

The vehicle fuel consumption is shown in Figure 11 for both UDDS and HWFET driving cycles. The results illustrate the advantage of the multi-mode LTC-SI engine in the mild hybrid over the single-mode SI. However, this improvement rate is smaller over the HWFET driving cycle since the engine operating points are located mainly in fuel-efficient regions independent of the electrification level.

Table 6 lists a detailed comparison in different engine electrification levels for the HWFET cycle. As expected, the fuel economy benefit of multi-mode LTC-SI is not substantial in the HWFET cycle where the fuel consumption improves 0.4% in the mild hybrid compared to the single-mode SI engine. This is due to less opportunity for the hybridization in the highway cycles to use the e-motor torque assist to shift the low efficient engine operating points to high efficient LTC regions at low/mid engine loads.

Figure 12 depicts the powertrain running modes for different hybridization levels for both UDDS and HWFET driving cycles. The x-axis shows the battery power and the y-axis shows the power demand at wheels. The regenerative braking happens once the power sign is negative (i.e., region III). Figure 12-[a.sub.2] and [b.sub.2] show that, in the mild hybrid, the x-axis range is smaller for the given demanded wheel power since the battery available power is smaller. Note that this figure aims to show the difference of engine usage for the PHEV and the mild hybrid and the size of a region does not show the amount of time the powertrain has been operating in that region. In the mild hybrid and high wheel demanded power, the engine operates in the SI mode (i.e., region II), whereas in the PHEV the engine does not run in the SI mode during the battery depletion (i.e., region II). This is because in the PHEV, the control strategy uses the engine in SI mode to charge the battery faster (i.e., regions I and II) since the battery has larger power limit compared to the mild hybrid. For the highway cycle the engine does not operate in the SI mode for both PHEV and mild hybrid. This is because that the demanded wheel power is less compared to UDDS and the engine only operates in SI mode to charge the battery (i.e., regions I and III) as well as to provide the power demand.

SUMMARY AND CONCLUSION

Fuel saving potential of utilizing a multi-mode LTC-SI engine in a parallel HEV was investigated. The multi-mode engine includes HCCI, RCCI, and SI modes. A hybrid electric setup was designed and built to provide experimental data for the model extraction and validation. The PMP optimal energy management control technique was designed for the LTC-SI HEV powertrain and analyzed for different hybridization level. The powertrain controller is designed to enable switching among different modes, while minimizing fuel consumption by including penalty for transient engine operations. In the parallel architecture, a clutch is added to the engine shaft to enable the only electric mode. Below are the main findings based on the optimization results in this study:

* The results for the UDDS cycle show the multi-mode LTC-SI engine offers up to 8.8% fuel saving over a single-mode SI engine in the parallel HEV. This improvement reduces to 1.8% for the HWFET driving cycle. This is because the engine spends less time in the LTC modes in the HWFET cycle, which leads to less opportunity to save fuel.

* The engine LTC running time increases from 1.5% to 32.8% by reducing the vehicle electrification level from PHEV to mild HEV, over the UDDS cycle. Moreover, the LTC running time changes from 0% in PHEV to 13.6% in the mild hybrid for the HWFET cycle. Among the LTC modes, the RCCI was the dominant mode with the most engine running time.

* In the parallel multi-mode LTC-SI powertrain, two ways for fuel savings include operating in the LTC mode or using electric torque assist offered by the e-motor. Fuel saving from these two ways are not additive. In strong hybridizations (e.g., PHEV) the optimizer commands to operate the engine in high power region of SI engine and turns off the engine for a longer time, while in mild hybrids, due to lack of sufficient electric power for charge-sustaining, the optimizer cannot locate the engine operating points in high power region, and, instead, the engine operating points shift to lower power regions which happen to be the LTC modes to save fuel.

* Compared to strong electrified vehicles such as PHEVs, low electrified vehicles such as mild hybrids are better suited to improve fuel economy in the multi-mode LTC-SI engine.

REFERENCES

[1.] Center for Climate and Energy Solutions. http://www.c2es.org/energy/source/oil. Accessed on 9 December 2016.

[2.] European Commission Climate Action. urlhttp://ec.europa.eu/clima/policies. Accessed on 15 December 2015.

[3.] Reitz R. and Duraisamy G.. Review of High Efficiency and Clean Reactivity Controlled Compression Ignition (RCCI) Combustion in Internal Combustion Engines. Progress in Energy and Combustion Science, 46(4):12-71, 2015.

[4.] McCarthy T.. Systems Approach to Making Optimal Powertrain Technology Choices - A Global OEM's Perspective. ASME ICE Fall Tech. Conference, Dear-born, MI, USA, 2013.

[5.] National Research Council. Cost, Effectiveness, and Deployment of Fuel Economy Technologies for Light-Duty Vehicles. Washington, DC: The National Academies Press, page 11, 2015. doi:10.17226/21744.

[6.] Kermani S., Delprat S., Guerra T., Trigui R., and Jeanneret B.. Predictive Energy Management for Hybrid Vehicle. Control Engineering Practice, 20(4):408-420, 2012.

[7.] Opila D.F., Wang X., McGee R., Gillespie R.B., Cook J., and Grizzle J.W.. An Energy Management Controller to Optimally Trade Off Fuel Economy and Drivability for Hybrid Vehicles. IEEE Transactions on Control Systems Technology, 20(6):1490-1506, 2012.

[8.] Johri R., Salvi A., and Filipi Z.. Optimal Energy Management For a Hybrid Vehicle Using Neuro-Dynamic Programming to Consider Transient Engine Operation. ASME Dynamic Systems and Control Conference, Paper No. 2011-6138, Oct. 31-Nov. 2, 2011, Arlington, Virginia, USA.

[9.] Jalil N. and Salman M.. A Rule-Based Energy Management Strategy for a Series Hybrid Vehicle. Proceedings of American Control Conference, 1:689-693, 1997.

[10.] Albert I.J., Kahrimanovic E., and Emadi A.. Diesel Sport Utility Vehicles with Hybrid Electric Drive Trains. IEEE Transactions on Vehicular Technology, 53(4):1247-1256, 2004.

[11.] Sciarretta A., Back M., and Guzzella L.. Optimal Control of Parallel Hybrid Electric Vehicles. IEEE Transactions on Control Systems Technology, 12(3):352-363, 2004.

[12.] Lin C., Peng H., and Grizzle J.W.. A Stochastic Control Strategy for Hybrid Electric Vehicles. Proceeding of the American Control Conference, 5:4710-4715, 2004.

[13.] Brahma A., Guezennec Y., and Rizzoni G.. Optimal Energy Management in Series Hybrid Electric Vehicles. Proceeding of the American Control Conference, 1(6):60-64, 2000.

[14.] Ferdowsi F., Yazdankhah A.S., and Kojabadi H.M.. An Optimized Combination of a Large Grid Connected PV System along with Battery Cells and a Diesel Generator. arXiv preprint arXiv:1604.06691, 2016.

[15.] Brahma A., Glenn B., Guezennec Y., Miller T., Rizzoni G., and Washington G.. Modeling, Performance Analysis and Control Design of a Hybrid Sport-Utility Vehicle. Proceeding of the IEEE Int. Conference on Control Application, 1:448-453, 1999.

[16.] Bassett M., Brods B., Hall J., Borman S., Grove M., and Reader S.. GPS Based Energy Management Control for Plug-in Hybrid Vehicles. SAE Technical Paper No. 2015-01-1226, 2015.

[17.] Sciarretta A., Serrao L., Dewangan P.C., Tona P., Bergshoeff E., Bordons C., Charmpa L., Elbert P., Eriksson L., and Hofman T.. A Control Benchmark on The Energy Management of a Plug-In Hybrid Electric Vehicle. Control Engineering Practice, 29:287-298, 2014.

[18.] Tribioli L., Barbieri M., Capata R., Sciubba E., Jannelli E., and Bella G.. A Real Time Energy Management Strategy for Plug-in Hybrid Electric Vehicles Based on Optimal Control Theory. Energy Procedia, 45:949-958, 2014.

[19.] Bassett M., Thatcher I., Bisordi A., Hall J., Fraser N., and Warth M... Design of a Dedicated Range Extender Engine. SAE Technical Paper No. 2011-01-0862, 2011.

[20.] Kessels J.T.B.A., Koot M.W.T., Van den Bosch P.P.J., and Kok D.B.. Online Energy Management for Hybrid Electric Vehicles. IEEE Transactions on Vehicular Technology, 57(6):3428-3440, 2008.

[21.] Rezaei A. and Burl J.B.. Effects of time horizon on model predictive control for hybrid electric vehicles. IFAC-PapersOnLine, 48(15):252-256, 2015.

[22.] Kawamoto, N., Naiki, K., Kawai, T., Shikida, T. ., "Development of New 1.8-Liter Engine for Hybrid Vehicles," SAE Technical Paper 2009-01-1061, 2009, doi:10.4271/2009-01-1061.

[23.] Yonekawa, A., Ueno, M., Watanabe, O., and Ishikawa, N., "Development of New Gasoline Engine for ACCORD Plug-in Hybrid," SAE Technical Paper 2013-01-1738, 2013, doi:10.4271/2013-01-1738.

[24.] Delorme A., Rouseau A., Wallner T., Ortiz-Soto E., Babajimopolous A., and Assanis D.. Evaluation of Homogeneous Charge Compression Ignition (HCCI) Engine Fuel Savings for Various Electric Drive Powertrains. 25th World Battery, Hybrid and Fuel Cell Electric Vehicle Symposium & Exhibition, China, 2010.

[25.] Lawler, B., Ortiz-Soto, E., Gupta, R., Peng, H. ., "Hybrid Electric Vehicle Powertrain and Control Strategy Optimization to Maximize the Synergy with a Gasoline HCCI Engine," SAE Int. J. Engines 4(1):1115-1126, 2011, doi:10.4271/2011-01-0888.

[26.] Solouk A., Shahbakhti M., and Mahjoub M.J.. Energy Management and Control of a Hybrid Electric Vehicle with an Integrated Low Temperature Combustion (LTC) Engine. ASME Dynamic Systems and Control Conference, Paper No. DSCC2014-6286, 22-24 Oct. 2014, San Antonio, TX, USA.

[27.] Solouk A. and Shahbakhti M.. Modeling and Energy Management of an HCCI based Powertrain for Series Hybrid and Extended Range Electric Vehicles. Int. Journal of Powertrains, 6(2). doi:10.1504/IJPT.2017.10001761, 2017.

[28.] Hanson, R., Spannbauer, S., Gross, C., Reitz, R. ., "Highway Fuel Economy Testing of an RCCI Series Hybrid Vehicle," SAE Technical Paper 2015-01-0837, 2015, doi:10.4271/2015-01-0837.

[29.] Solouk A. and Shahbakhti M.. Energy Optimization and Fuel Economy Investigation of a Series Hybrid Electric Vehicle Integrated with Diesel/RCCI Engines. Energies, 9(12). doi:10.3390/en9121020, 2016.

[30.] Solouk, A., Shakiba-herfeh, M., K. Kannan, H., Solmaz, H.., "Fuel Economy Benefits of Integrating a Multi-Mode Low Temperature Combustion (LTC) Engine in a Series Extended Range Electric Powertrain," SAE Technical Paper 2016-01-2361, 2016, doi:10.4271/2016-01-2361.

[31.] Kim N., Cha S., and Peng H.. Optimal Control of Hybrid Electric Vehicles Based on Pontryagin's Minimum Principle. IEEE Transactions on Control Systems Technology, 19(5):1279-1287, 2011.

[32.] Nuesch S. P.. Analysis and Control of Multimode Combustion Switching Sequence. PhD Thesis, The University of Michigan, 2015.

[33.] Yazdani, A. and Bidarvatan, M., "A Comparative Analysis for Optimal Control of Power Split in a Fuel Cell Hybrid Electric Vehicle," SAE Technical Paper 2016-01-1189, 2016, doi:10.4271/2016-01-1189.

CONTACT INFORMATION

Ali Solouk

asoloukm@mtu.edu

Dr. Mahdi Shahbakhti

mahdish@mtu.edu

ACKNOWLEDGMENTS

This work was partially supported by the United States National Science Foundation under Grant No. 1434273. LG Chem Power Inc. is gratefully acknowledged for providing the battery packs in this study. Authors would like to thank Dr. Seyfi Polat, Dr. Hamit Solmaz, Paul Dice, and Michigan Tech. Energy Mechatronis Laboratory (EML) graduate students who assisted us with building the experimental powertrain setup in this study.

NOMENCLATURE

ABBREVIATIONS

BSFC - Brake specific fuel consumption

BTE - Brake thermal efficiency

CI - Compression ignition

EV - Electric vehicle

EVC - Exhaust valve closing

EREV - Extended range electric vehicle

HEV - Hybrid electric vehicle

HCCI - Homogeneous charge compression ignition

HWFET - Highway fuel economy test

ICE - Internal combustion engine

IVO - Intake valve opening

LTC - Low temperature combustion

LVD - Longitudinal vehicle dynamic

OCV - Open circuit voltage

PMP - Pontryagin minimum principle

RCCI - Reactivity controlled compression ignition

RON - Research octane number

SOC - State of charge

SOI - Start of ignition

SI - Spark ignition

UDDS - Urban dynamometer driving schedule

SYMBOLS

[lambda] - Co-state [-]

[C.sub.d] - Vehicle aerodynamic drag coefficient [-]

[[??].sub.f] - Fuel consumption rate [g/sec]

[f.sub.r] - Rolling resistance coefficient [-]

[omega] - Engine speed [rpm]

[[omega].sub.motor] - E-motor speed [rpm]

A - Vehicle frontal area [[m.sup.2]]

H - Hamiltonian [g/sec]

M - Vehicle total mass [kg]

[n.sub.c] - mechanical coupling ration [-]

[n.sub.t] - transmission ration [-]

[n.sub.d] - differential ratio [-]

[P.sub.Motor,mech] - Motor traction mechanical power [kW]

[P.sub.Motor,e] - Motor regen electrical power [kW]

[P.sub.bat] - Battery power [kW]

[P.sub.wheel] - Driver power demand [kW]

[P.sub.eng] - Engine generated power [kW]

[Q.sub.nom] - Battery nominal energy capacity [Wh]

R - Battery internal resistance [[ohm]]

[V.sub.veh] - Vehicle speed [[m/sec]]

r - Wheel radius [m]

p - Air density [[kg/[m.sup.3]]]

[theta] - Road slope [[degrees]]

[lambda] - Co-state

[m.sub.ij] - Mode-switching fuel penalty [g]

u - Control variable

SUBSCRIPTS

bat - Battery

drag - Air drag

eng - Engine

e - Electrical

grade - Gradeability

gear - Gearbox

intake - Engine intake

min - Minimum

max - Maximum

motor - Electric motor

mech - Mechanical

nom - Nominal

veh - Vehicle

roll - Rolling resistance

APPENDIX

LVD MODEL DESCRIPTION

The purpose of the LVD model is to calculate the vehicle actual speed ([V.sub.veh]) based on vehicle dynamics [33]:

[MV.sub.veh] [[dV.sub.veh]/dt] = [P.sub.wheel] - [P.sub.roll] - [P.sub.drag] - P.sub.gdrag] (18)

where M is the vehicle total mass; [P.sub.drag], [P.sub.roll], and [P.sub.grade] are aerodynamic drag, rolling resistance and gravity powers, respectively. These parasitic powers are calculated by:

[P.sub.roll] = [M.sub.gfr](1 + [[V.sub.veh]/100])[V.sub.veh] (19)

[P.sub.drag] = [[1/2][rho]A[C.sub.d][V.sub.veh].sup.3] (20)

[P.sub.grade] = [MV.sub.vehg] sin [theta] (21)

where [rho] is the air density, A is the vehicle frontal area, [C.sub.d] is the vehicle aerodynamic drag coefficient, [f.sub.r] is the rolling resistance coefficient, and [theta] is the road slope.

Ali Solouk

Michigan Technological University

Mohammad Shakiba-herfeh

Ford Motor Company

Mahdi Shahbakhti

Michigan Technological University

doi:10.4271/2017-01-1153
Table 1. Parameters of the baseline engine in this study

Parameters                  Value/Description

Engine Model                GM Ecotec LHU
Bore X Stroke                86 X 86 mm
Number of Cylinders           4
Displacement Volume           2.0 L
Compression Ratio             9.2:1
Connecting Rod Length       145.5 mm
Max Power                   270 hp@6000 rpm
Fuel Injection System       Gasoline Direct Injection
Valve System                DOHC 4 Valves

Table 2. Engine Operating Conditions

Combustion Mode          SI            HCCI

Injection Rail Used      DI            PFI1+PFI2
Fuel Type                Gasoline      n-heptane+ iso-octane
Engine Speed (rpm)       800-4000       800-1600
SOI (CAD bTDC)           100            450
IVO (CAD bTDC)           -24.5           25.5
EVC (CAD bTDC)            22             22
[T.sub.intake]
([degrees]C)              40             40,60,80,100
[P.sub.intake](kPa)      100            100
RON (-)                   87              0-[40.sup.1]

Combustion Mode          RCCI

Injection Rail Used      DI+PFI1
Fuel Type                n-heptane + iso-octane
Engine Speed (rpm)        800-2200
SOI (CAD bTDC)            Variable
IVO (CAD bTDC)             25.5
EVC (CAD bTDC)             22
[T.sub.intake]
([degrees]C)               40,40,80,100
[P.sub.intake](kPa)       100
RON (-)                    20-[60.sup.1]

Table 3. Vehicle specifications.

Parameters                                    Values

Vehicle Curb Mass, M                          1775 (kg)
Frontal Area, A                                  2.0 ([m.sup.2])
Engine Motor Coupling Gear Ratio, [n.sub.c]      2.0
6-Speed Transmission Gear Ratios                [3.166, 2.05, 1.481,
                                                 1.166, 0.916, 0.725]
Differential Ratio, [n.sub.d]                    4.529
Wheel Radius, r                                  0.33 (m)
Drag Coefficient, [C.sub.d]                      0.25
Rolling Resistance Coefficient, [f.sub.r]        0.01

Table 4. Hybridization level definition in this study

Hybridization   [[P.sub.bat]/[P.sub.eng]]   Electric Motor Power (kW)

PHEV                   1.0                         60
Full Hybrid            0.65                        40
Mild Hybrid            0.30                        18

Hybridization     Operating Voltage (V)

PHEV                     270-410
Full Hybrid              180-270
Mild Hybrid               80-120

Table 5. Results for both multi-mode LTC-SI and single-mode SI engine
in different electrification levels during UDDS driving cycle

                                  Multi-Mode LTC-SI
Metrics                PHEV      Full Hybrid    Mild Hybrid

Fuel consumption (g)   352.6        347.9        369.4
Ave. engine BTE (%)     34.2         34.9         32.8
Engine work (MJ)         5.26         5.26         5.31
Engine ON time (sec)   179          340          537
Battery loss (kJ)      617.5        440.4        437.2

                                Single-Mode SI
Metrics                PHEV    Full Hybrid   Mild Hybrid

Fuel consumption (g)   357.6     370.0         405.0
Ave. engine BTE (%)     34.2      33.6          30.8
Engine work (MJ)         5.33      5.48          5.50
Engine ON time (sec)   172       320           573
Battery loss (kJ)      653.4     452.2         541.0

Table 6. Results for both multi-mode LTC-SI and single-mode SI engine
in different electrification levels during HWFET driving cycle

                                   Multi-Mode LTC-SI
Metrics                  PHEV          Full Hybrid      Mild Hybrid

Fuel consumption (g)     1162.1        1181.7           1210.0
Ave. engine BTE (%)        34.4          34.3             32.6
Engine work (MJ)           17.4          17.6             17.6
Engine ON time (sec)      676           870              865
Battery loss (kJ)        1084.4         898.9            805.9

                                  Single-Mode SI
Metrics                  PHEV       Full Hybrid      Mild Hybrid

Fuel consumption (g)     1162.1        1181.3           1215.0
Ave. engine BTE (%)        34.2          34.3             32.2
Engine work (MJ)           17.4          17.6             17.8
Engine ON time (sec)      676           813              850
Battery loss (kJ)        1124.1        1027.7            855.9
COPYRIGHT 2017 SAE International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Solouk, Ali; Shakiba-herfeh, Mohammad; Shahbakhti, Mahdi
Publication:SAE International Journal of Alternative Powertrains
Article Type:Report
Date:May 1, 2017
Words:8013
Previous Article:Seasonality Effect on Electric Vehicle Miles Traveled in Electrified Vehicles.
Next Article:Powersplit or Parallel - Selecting the Right Hybrid Architecture.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters