Professional Documents
Culture Documents
ABSTRACT The control accuracies of the injection time and fuel quantity are the most important factors
affecting the fuel economy and emissions of an engine equipped with a high-pressure common-rail fuel
injection system. This paper presents an intelligent dual-voltage driving control that can simultaneously
reduce the response time of the injector and improve the accuracy of the injected fuel quantity. The designed
method employs a novel boost dc–dc circuit and does not require an additional independent dc–dc module
that includes an inductor switch MOSFET and a controller. During the time interval between two continuous
injections, the low-side MOSFET serves as a dc–dc switch, and the software in the controller controls the
dual-voltage dc–dc, which uses the injector as a charging inductor of the dc–dc module. The operation
principle of this intelligent driving circuit is described. Experimental results show that this method can not
only significantly decrease the response time of the injector but also improve the stability of the fuel injection
process.
INDEX TERMS Intelligent driving circuit, dual-voltages driving control, common-rail injector,
dc-dc charger, fuel injection control.
2169-3536
2017 IEEE. Translations and content mining are permitted for academic research only.
VOLUME 5, 2017 Personal use is also permitted, but republication/redistribution requires IEEE permission. 27681
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
J. Xiong, H. Gu: Intelligent Dual-Voltage Driving Method and Circuit for a Common Rail Injector
IC is required for the high-side solenoid drive (Sakamoto K, II. WORKING PRINCIPLE OF THE
Nunogawa Y, 1999) [8]. Delphi Co., Ltd. uses this driving COMMON-RAIL INJECTOR
circuit. However, there are two shortcomings of this circuit. When the electronic control unit of a high-pressure common-
The first is that the driving voltage varies with the battery, rail diesel engine is powered on and the injector starts to
which can cause large deviations in the fuel quantity. The inject fuel into the combustion chamber, the rate of increase
second is that the response time of the injector is slow and of the current of the solenoid valve should be large enough
unstable. until the current reaches the peak so that the armature of the
To achieve flexible injection rates in a common-rail solenoid valve can be pulled up quickly. When the armature
system, we need to improve the response time of the is attracted to the coil, the air gap of the magnetic path is
injector to realize multiple injections during a short time. decreased; thus, the magnetic reluctance drops. Afterwards,
A rapid response by the fuel injector is very impor- the solenoid valve only requires a small holding current,
tant. The factors determining the response time include which can reliably maintain the armature at the attracted
the mechanical design of the injectors (Dean C, Irena C, position. When the solenoid valve is closed, a smaller holding
Aleksandar S. et al. 2008) [9], the electromagnetic design current is imposed to accelerate the later action speed. Mean-
of the fast speed solenoid coil, and the optimization of the while, such an injection control method can decrease both the
driving circuit (Hideyuki W et al. 2005) [10]–[12]. Kong and electric energy consumption of the battery and the heating
Li (2014) replaced the single coil of a high-speed solenoid loss of the coil in the solenoid valve. In addition, fuel leakage
valve by parallel coils and found that both the delay time and occurs between the needle valve and the plunger control. The
the switching time were decreased greatly [13]. In general, confluence of the leaking fuel and the return fuel from the
the response time of the solenoid valve inside the fuel injec- high-pressure fuel supply pump or the pressure control valve
tor is approximately 200∼250 µs. For heavy-duty engines, flows back to the fuel tank through the fuel return pipe.
injectors with solenoid valves are widely adopted at present When a diesel engine is running, the high-pressure fuel
due to their high reliabilities and low costs. supply pump delivers fuel continuously to the common-rail,
and a high pressure is established in the rail. The work- where U is the coil voltage, i is the solenoid current, R is the
ing process of a traditional single-voltage driving injector is total resistance of the solenoid coil, w is the total magnetic
illustrated in the following five stages. The corresponding flux of the coil, and t is the time.
variation profiles of the main physical parameters are shown Supposing the armature position is fixed and the magnetic
in Figure 4. The five stages in one injection process are: circuit is not saturated, the coil inductance can be considered
a constant. Thus, the total magnetic flux is denoted by
• 0: The injector is closed.
• 1: The current increases to the peak (start of injection). w = Li (2)
• 2: The armature begins to lift (start of injection).
• 3: The armature moves up. The opening speed of the
And
nozzle needle valve depends on the fuel flow velocities U = iR + L dw (3)
dt
into the control chamber and out of the fuel drain hole.
When the piston valve reaches the top position, the The solution of this equation is
nozzle needle is fully opened and the fuel is injected into t
the combustion chamber with a very high pressure. i = Iw (1 − e− T ) (4)
• 4: The armature moves down to the seated position.
where Iw is a steady-state current of the coil and Iw = U /R;
If the solenoid valve is not energized, the armature moves T is the constant time of the solenoid coil, which is related to
downward by the acting force of the armature spring. Finally, the rising rate and the initial current and T = L/R.
the armature pushes the pivot ball to seal the fuel drain hole. Let i = Icd . Substitute this into Eq. (4) and taking the eddy
Once the fuel drain hole is closed, the fuel from the high- current effect into account, a coefficient α is used (generally,
pressure common-rail flows into the control chamber where α is within 1.1∼1.3.), and thus, the attracting time is obtained.
a high pressure is built up accordingly. This high pressure is
Iw
almost consistent with that of the common rail. As a result, t01 = αT ln (5)
an increase in the force acting on the end surface of the Iw -Icd
valve piston is generated. If the summation of this force If the armature starts to move, the magnetic flux increases
and the nozzle needle spring force is greater than the fuel because this movement causes the air gap to gradually
force generated by the bearing cavity, the nozzle needle valve decrease. The change in the magnetic flux generates a
is closed. The closing speed of the nozzle needle depends counter electromotive force (EMF) in the coil. The generated
on the velocity that the fuel flows into the control chamber. EMF and the self-induced EMF prevent current growth in
the coil when the coil inductance varies. At this moment, the
voltage balance equation of the solenoid coil and the motion
equation of the armature are expressed by
U = iR + L dw dL
dt + i dt (6)
2
Fd − Ff = m ddt 2x (7)
where Fd is the electromagnetic force, Ff is the electromag-
net reaction force, m is the total mass of the armature and
other moving parts, and x is the displacement of the armature.
These two forces can be measured on an injector testing
bench. The first term on the right of the voltage balance
equation is the self-induction electromotive force. The second
term is the back electromotive force when the inductance FIGURE 6. Dual-voltage driving method.
changes. As a result, the current growth time is less than The corresponding current and voltage of the solenoid valve
that of an exponential function. Then, the action time of the using a PWM modulated signal are shown in Figure 6. In the
solenoid valve can be determined by single-voltage driving method, the driving voltage is the bat-
q tery voltage, which can be seen from the traditional driving
t02 = F2mx d
(8)
d −Ff schematic shown in Figure 2.
Supposed that the battery voltage is steady at 24 V, the driv-
ing current and voltage are shown by real lines in Figure 5.
However, in practical working conditions, the battery voltage
will vary with the capacity or the working temperature. Thus,
the attracting time t01 will not be kept at a fixed time and
will results in an unstable response time of the injector. The
situation is denoted by the dashed lines in Figure 5.
C. DUAL-VOLTAGE DRIVING CIRCUIT
If the driving voltage in the first stage is increased from
24 V to 48 V and the current rises to the same peak value
of 25 Amperes, the energizing time will be much shorter
than that with 24 V. The response time of the first stage will
decrease from 239.40 µs to 110.5 µs. The driving current
and voltage of the dual-voltage driving method are shown in
FIGURE 5. Traditional single-voltage driving method. Figure 6.
The normal dual-voltage driving circuit is shown in
B. NORMAL SINGLE-VOLTAGE DRIVING CIRCUIT
Figure 7, where an independent DC-DC boost module is used
Although the structure and working principle of the common-
to set up the high voltage. The DC-DC driving circuit consists
rail injector is similar, the response time and the fuel quantity
of a DC-DC controller (U1), a switch MOSFET Q7, a diode
precision of the designed dual-voltage driving method are
D1, and a boost capacitor C1. By using a DC-DC booster, a
quite different from that of the single-voltage driver because
more stable and regular boosting high voltage can be obtained
there are many mechanical and electronic parameters that
no matter what the battery voltage is. Thus, the response time
can affect the accuracy of the fuel injection quantity and its
and the accuracy of injected fuel quantity will not be affected
response time. In this study, we first analyze the response time
by the variation of the battery voltage. However, this normal
and the precision of the single-voltage driving method. The
driving circuit requires some more electronic components and
traditional single-voltage driving approach for a common-rail
a larger hardware area of the ECU, which will increase the
injector is shown in Figure 5, and the corresponding physical
cost and decrease the system reliability. In this study, a new
stages are already described in Figure 4.
dual-voltage driving method is introduced that does not need
As measured on an injector test bench, the electromagnetic
the additional inductor and switch shown in the red dashed
force Fd for a designed common-rail injector is 78 N, while
box in Figure 7.
the electromagnet reaction force Ff is 50 N. The inductance
of the injector is measured as 165 µH, and the resistance is IV. DESIGNED INGELLIGENT DUAL-VOLTAGE
0.2531 . The driving voltage during the attracting time is DRIVING CONTROL
24 V, so according to Eq. (4) Iw = 98.42 A. Meanwhile, A. DUAL-VOLTAGE SELF-CHARGING DRIVING CIRCUIT
according to Eq. (4), T=916.6 µs. Substituting these values WITH NO ADDITIONAL DC-DC BOOST CONVERTER
into Eq. (5) and setting Icd to 25 A, the attracting time In this study, an intelligent dual-voltage driving circuit is
t01 is 239.4 µs where α is set to an average value of 1.2. designed. Here, the inductor of the typical DC-DC boost
FIGURE 7. Normal dual-voltage driving circuit. FIGURE 9. Designed non-independent DC-DC boost circuit
B. DC-DC MODEL
As shown in Figure 9, taking the injection process of the
first cylinder as an example, during the injection time gap
between two cylinders, the high-side boost voltage switch
is shut down, and the high-side PWM MOSFET is open.
Accordingly, the current flow path is shown as the red lines in
Figure 9. Under this working condition, the equivalent circuit
model of the DC-DC boost converter is shown in Figure 10,
where it is obvious that those components composed a typical
FIGURE 8. Designed dual-voltage driving circuit. boost converter.
At this time, energy is being stored in the inductor L, which Mode 2 (S=1, D=0):
is also the injector in the designed intelligent dual-voltage
driving circuit. The working process can be divided into four
S=1
modes depending on the state of the diode and the switch d = 0
(THEUNISSE T A F, 2015). Considering the switching ele- d 1 (10)
dt vc = − Rc vc
ments (D is an idea diode and S is the low-side switch Q1),
d iL = − E
the equations of these four modes are: dt L
Mode 1 (S=0, D=1): Mode 3 (S=0, D=0):
S=0
S=0
d = 1
d = 0
d 1 1 (9) d 1 (11)
dt vc = − Rc vc + c iL
dt vc = − Rc vc
d iL = − 1 vc + E
d iL = 0
dt L L dt
Mode 4 (S=1, D=1): where D is the duty cycle, D = ton /(ton + toff ), and fsw is the
frequency of the switch fsw = 1/T .
S=1
1Uout = ton Iout
d = 1
Cout (15)
d (12) During the second half cycle, the power FET Q1 turns off,
dt vc = 0
d iL = E
and the Schottky diode D1 is forward biased, as shown in
dt L Figure 12 (c). The energy stored in the inductor is pumped
out to charge the output capacitor. The Schottky diode side
where S=0 indicates that Q1 is switched off, and D=0 indi- of the inductor is clamped to above the output voltage by a
cates that the diode is shut off. R is the impedance of the Schottky diode. Therefore, the voltage decrease across the
injector solenoid. inductor equals Uin −Uout . The change in the inductor current
Combining the operation conditions together with the dif- during the second cycle is
ferential equations for each mode according to the state of the
toff (Uin −Uout )
switch, it is clear that Mode 2 and Mode 4 can be combined 1I2 = L (16)
into a single mode with the same dynamics, and Mode 1 and 1toff = (1−D)
(17)
Mode 3 can also be combined into a single mode; these modes fsw
are shown in Figure 12 (b) and (c), respectively. For the steady operation conditions, the same amount of
energy stored in the inductor must be taken out. The change
in inductor current during the two cycles must be the same.
1I1 +1I2 = 0 (18)
Substituting Eqs. (13) and (16) into (18), we get
Uout −Uin
D= Uout (19)
For example, if the boost voltage Uout is set to 48 V and
the input voltage Uin is 24 V, according to Eq. (19), D is
50%. At this time, the MCU outputs a PWM signal whose
duty is 50% to control the MOSFET cyc1e. Normally, the
Uout is a constant value, but the Uin is a varied value, so the
MCU should sample the input voltage to adjust the PWM
duty, and a higher accuracy can be obtained by this closed-
loop control method. On the other hand, the current of the
FIGURE 13. Designed ECU hardware including the intelligent driver. recharger should be lower than the hold current of 12 A.
According to Eq. (5), the attracting time t01 is 110.58 µs,
When the MOSFET Q1 stops conducting, the inductor where Iw equals to 94.82 A and Icd is 12 A.
voltage flies back or reverses because the current through the If we want to limit the maximum DC-DC recharge current,
inductor cannot change instantaneously. The voltage across such as lower than the hold current of 12 A, we must control
the inductor increases to a value that is higher than the the current response time to less than 1058 ms, which requires
combined voltage across the diode and the output capacitor. the recharge frequency to be greater than 9447.6 Hz. Taking
As soon as this value is reached, the diode starts conduct- into account the influences of the ripple wave and the boost
ing, and the voltage that appears across the output capacitor capacitor, the recharge frequency is set to 20 kHz.
is higher than the input voltage. In Figure 13, the switch
MOSFET is controlled by a PWM signal to control the boost C. ECU HARDWARE DESIGN
chopper, where the maximum PWM duty must be limited. According to the designed dual-voltage driving circuit, a
During the first half cycle shown in Figure 12 (b), the real ECU was developed in which the microprocessor is an
internal power FET Q1 turns on, and the Schottky diode D1 MPC5634M chip manufactured by Freescale Semiconductor
is reversely biased and cuts off the current flow to the output. Company. It has 96 KB RAM and 2 MB FLASH. The devel-
The output current is supplied from the output capacitor. oped ECU has 16 channels of analog signal input, 20 channels
If the voltage across the inductor (injector) is Uin , the inductor of digital signal input, 1 channel for the crank shaft posi-
current ramps up at a rate of Uin /L and L is the inductance tion sensor signal input, 1 channel for the camshaft position
of the injector. The inductance is magnetized, and energy is sensor signal input, 1 channel for the vehicle rotation speed
stored in the inductor. The change in the inductor current is signal input, 1 channel for the air flow frequency input. The
obtained by main output channels include 6-cylinder injection controls
(peak injection current > 25 Amperes, holding current >
1I1 = ton Uin
L (13) 15 Amperes - these currents can be adjusted and support
D
ton = fsw (14) at least 5 injections for each injection process), 1 channel
for the EGR valve driver, 1 channel for the engine rotation
speed output, and 2 channels for the H-bridge driver with each
channel’s current greater than 5 Amperes. The printed circuit
board (PCB) of the designed ECU is shown in Figure 13.
V. EXPERIMENTAL RESULTS
To validate the feasibility of the designed dual-voltage driving
method, an experimental test was performed on an injector
test bench manufactured by French EFS Co., Ltd., which is
shown in Figure 14. An EMI2 flow meter was used to mea-
sure the instantaneous volume of the fuel quantity injected
stroke-by-stroke.
In the test, the boost voltage of the first stage was set to
48 V, according to Eq. (4), Iw was 189.65 Amperes, and T
was 651.9 µs. Then, substituting these two values into Eq. (5)
and setting Icd equal to 18 Amperes (this value depends on
the characteristics of the injector), the attracting time t01 was
determined to be 110.5 µs. The measured solenoid current
profile is shown in Figure 15. The measured currents and the
attracting time are very close to the estimated value, which
indicates that the designed dual-voltage driving method is
feasible. The response time of the first stage is shown in
Figure 15 (b) and (c). These figures show that under different
battery voltage conditions, the first stage response time is
the same. The injection rate profile with different battery
voltage conditions is shown in Figure 15 (d), which indicates
that under different battery voltage conditions, the injector
operates approximately the same.
To evaluate the injection performance improvement of the
designed dual-voltage driving method, more tests were con-
ducted on the EFS injector test bench. For each test, differ-
ent fuel pressures of the common rail and energizing times
(including the attracting time and the duration of the hold- FIGURE 15. Results of the solenoid current and injection rate of the
dual-voltage driving method. (a) Current Curve .(b) Response Time
ing current) were selected when the pump speed was set to (Battery voltage = 18 Volts) . (c) Response Time (Battery
700 RPM, and the corresponding injected fuel quantities were voltage = 24 Volts) .(d) Relevant Injection Rate.
recorded. In contrast, the results from a traditional single-
voltage driving method were also measured. The results of 800, 1200, and 1600 bar, respectively. The standard deviation
the measured standard deviation of the injected fuel quantities of the injected fuel quantity can be decreased significantly by
are listed in Tables 1 to 3. using the designed intelligent dual-voltage driving method,
In the tests, three typical work conditions of a diesel which means the cycle variation of diesel engines can be
engine were selected. The corresponding rail pressures were reduced, and a higher emission standard can be achieved.
TABLE 1. Standard deviation of the injected fuel quantity when the rail [6] J. Kashdan, P. Anselmi, and B. Walter, ‘‘Advanced injection strategies
pressure is 800 bar. for controlling low-temperature diesel combustion and emissions,’’ SAE
Tech. Paper 2009-01-1962, 2009.
[7] C. Beatrice, ‘‘Experimental analysis of the operating parameter influence
on the application of low temperature combustion in the modern diesel
engines,’’ SAE Tech. Paper 2007-01-1839, 2007.
[8] K. Sakamoto, Y. Nunogawa, K. Satonaka, T. Kouda, and S. Horiuchi,
‘‘An intelligent power IC with reverse battery protection for fast-switching
high-side solenoid drive,’’ IEEE Trans. Electron Devices, vol. 46, no. 8,
pp. 1775–1781, Aug. 1999.
[9] C. Dean, C. Irena, and S. Aleksandar, ‘‘Improved performance of the elec-
tromagnetic fuel injector solenoid actuator using a modelling approach,’’
Int. J. Appl. Electromagn. Mechan., vol. 27, no. 4, pp. 251–273, 2008.
TABLE 2. Standard deviation of the injected fuel quantity when the rail
pressure is 1200 bar.
[10] W. Hideyuki and I. Shinya, ‘‘Development of compact and high per-
formance fuel injector using electromagnetic field simulation,’’ SAE
Tech. Paper 2005-01-0019, 2005.
[11] M. Piron, P. Sangha, G. Reid, T. J. E. Miller, and D. M. Ionel, ‘‘Rapid
computer-aided design method for fast-acting solenoid actuators,’’ IEEE
Trans. Ind. Appl., vol. 35, no. 5, pp. 991–999, Sep. 1999.
[12] W. C. Tsai, Z. H. Wu, P. C. Yu, and S. B. Chang, ‘‘Design of electrical
driving circuits for a high-pressure fuel injector and control of fuel injection
quantities by using the polynomial curve fitting method,’’ Int. J. Intell. Inf.
Process., vol. 2, no. 2, pp. 60–73, 2011.
[13] X. Kong and S. Li, ‘‘Dynamic performance of high speed solenoid valve
with parallel coils,’’ Chin. J. Mech. Eng., vol. 27, no. 4, pp. 816–821, 2014.
TABLE 3. Standard deviation of the injected fuel quantity when the rail [14] X. Zhang, H. Ren, and S. Pyo, ‘‘A high-efficiency DC-DC boost converter
pressure is 1600 bar. for a miniaturized microbial fuel cell,’’ IEEE Trans. Power Electron.,
vol. 30, no. 4, pp. 2313–2327, Apr. 2015.
[15] F. Alonge and R. Rabbeni, ‘‘Identification and robust control of
a quadratic DC/DC boost converter by Hammerstein model,’’ IEEE
Trans. Ind. Appl., vol. 51, no. 5, pp. 3975–3985, Sep./Oct. 2015,
doi: 10.1109/TIA.2015.2416154.
[16] T. A. F. Theunisse, J. Chai, R. G. Sanfelice, and W. P. M. H. Heemels,
‘‘Robust global stabilization of the DC-DC boost converter via hybrid
control,’’ IEEE Trans. Circuits Syst. I, Reg. Papers, vol. 62, no. 4,
pp. 1052–1061, Apr. 2015.
[17] W. C. Tsai, ‘‘Development of a high-pressure injector driving circuit for
controlling the air-fuel ratio of GDI engines,’’ in Proc. Int. Symp. Comput.
Consumer Control, Taichung, Taiwan, Jun. 2014, pp. 1275–1278.
VI. CONCLUSIONS [18] J. Schaeuffele and T. Zurawka, ‘‘Automotive software engineering: Prin-
Compared with a traditional single-voltage driving control ciples, processes, methods, and tools,’’ SAE Int., Warrendale, PA, USA,
Tech. Rep., 2005.
method, the designed intelligent dual-voltage driving method [19] F. W. Yan, P. Zhang, and C. Q. Du, ‘‘Novel evaluation method of fuel
can not only accelerate the response time of the fuel injec- consumption and emission for heavy-duty hybrid electric vehicles,’’ Int.
tor significantly but can also improve the operation stabil- J. Autom. Technol., vol. 15, no. 5, pp. 773–779, Aug. 2014.
ity of the injector. In particular, the driving method in this
study designed a novel boost circuit, which needs neither a
DC-DC controller nor its switch MOSFET and the indepen-
dent inductor. The driving circuit presented in this paper can JIAN XIONG received the B.S and M.S. degrees in
further improve the reliability of the hardware and reduce sys- electronic engineering and computer Science from
tem cost. Moreover, the designed intelligent driving method UESTC in 1997 and 2005 respectively. He is cur-
rently pursuing the Ph.D. degree in control theory
can adapt its boost voltage by setting the corresponding oper- and control engineering with the Dalian University
ation parameters such as the peak current and the holding of Technology. His research interests include wire-
current in the control software to fit different injectors. less communication and automotive electronics.
REFERENCES
[1] T. V. Johnson, ‘‘Review of diesel emissions and control,’’
SAE Tech. Paper 2010-01-0301, 2010.
[2] T. Lee and R. D. Reitz, ‘‘The effects of split injection and swirl on a
HSDI diesel engine equipped with a common-rail injection system,’’ SAE
HONG GU received the bachelor’s and master’s
Tech. Paper 2003-01-0349, 2003.
degree from the Shengyang University of Technol-
[3] Y. M. Waseda and D. Tanaka, ‘‘The effect of intake, injection param-
eters and fuel properties on diesel combustion and emissions,’’ SAE
ogy in 1982 and 1984, respectively, and the Ph.D.
Tech. Paper 2003-01-1793, 2003. degree from Zhejiang University in 1990. He is
[4] P. Li, Y. Zhang, and T. Li, ‘‘Elimination of fuel pressure fluctuation and currently a Professor with the Dalian University of
multi-injection fuel mass deviation of high pressure common-rail fuel Technology. His research interests include control
injection system,’’ Chin. J. Mech. Eng., vol. 28, no. 2, pp. 294–306, 2015. theory and control engineering, machine learning,
[5] K. Funai, T. Yamaguchi, and S. Itoh, ‘‘Injection rate shaping technology big data analysis, and data mining.
with common rail fuel system (ECD-U2),’’ SAE Tech. Paper 960107,
1996.