DEVELOPMENT OF A NOVEL 5kW/42V INTELLIGENT

CONVERTER FOR AUTOMOTIVE APPLICATIONS

A. Bousbaine*, M. Eljarh Author *
*University of Derby, School of Technology, Derby, UK, [email protected]
*University of Derby, School of Technology, Derby, UK, [email protected]

Keywords: Interleaved, DC-DC converters, PWM controller, cut carbon emissions in transport by 60% by 2050. One of its
42V Power-Net, automotive. key goals is no more conventionally-fuelled cars in the cities.
In a world of rising oil prices, growing congestion and
Abstract looming climate change, the EU’s transport system needs a
radical overhaul to maintain its role of growth engine and
keep pace with mounting global competition. In order to
Growing pressure on the automotive industry to produce cars
avoid having to limit EU’s freedom of movement, EU needs
with less exhaust emission, better fuel economy, and to save
to break the shackles of transport’s dependency on oil, but
energy necessitated the introduction of higher voltage
without sacrificing its efficiency. Electric vehicles (EVs), due
electrical power system to meet these requirements in short to
to their zero local and potentially minor green house gas
mid-term. Already, various electric systems architectures
emissions, are considered the cleanest option and provide
have been proposed and investigated over the past ten years.
opportunities in terms of efficiency and flexibility in the use
To meet such growing demands, the automotive industry has
of energy. Additionally, and especially in urban areas, road
to move to higher voltage and the 42V power-net system is
transport is also one of the major sources of emissions which
the preferred option [8]. The 42V is further processed by the
are harmful to human health [3].
interleaved six-phase dc-to- dc buck converter system to
supply power to the conventional automotive loads that are
The current 14V bus and Lundell based alternators are being
expected to remain at 14V level as well as to absorb the peak
pushed to the limits to meet the higher demand from current
transients on the 42V bus voltage. A special DC/DC converter
luxury cars and this will be further stretched moving a fully
is therefore needed to interconnect the 14V and the 42V DC
electrical car. Industry consortia such as SICAN in Europe
buses in the car of the future. Automotive electronics place
and MIT in USA propose that a 42V or higher voltage bus
severe demands on the performance and price of power
system for next generation cars is a must
electronic components and making the development of a
suitable converter a challenging task. This paper will present
The hybrid vehicles achieve about 100% improvement in the
the initial development of the 5kW/42V intelligent converter
fuel efficiency compared with vehicles using the similar fuel
for automotive applications using Matlab/Simulink.
engine displacement [10]. Previous and current research
conducted at the University of Derby [8] highlighted that a
1 Introduction centralised single battery system is an ideal option forward.
Growing concerns about public health, global warming and Figure 1 shows a complete Simulink model of the proposed
system.
economic safety are calling for sustainable road transport
technologies. Road transport contributes about one-fifth of the
The multileveled DC-to-DC converter is considered to be the
EU's total emissions of carbon dioxide (CO2), the main
greenhouse gas. While emissions from other sectors are heart of the system as it reduces 42V to 14V to be used by the
generally falling, those from road transport have continued to lower voltage loads. The converter architecture is shown in
increase since 1990. Passenger cars alone are responsible for figure 2.
around 12% of total EU emissions of carbon dioxide (CO2),
the main greenhouse gas. Due to those developments a compact, high efficiency dc-dc
converter is needed to enable the introduction of new
As part of the EU's efforts to tackle climate change, the modifications to the overall automobile design. The
European Commission proposed legislation, back in 2007, performance of the dc-dc converter will be integral part of the
setting emission performance standards for new passenger 42/14V dual bus system, and will help reduce costs and
cars. More recently white paper 2011, adopted a roadmap of weight for future automobiles.
40 concrete initiatives for the next decade to build a
competitive transport system that will increase mobility, The dc-dc converter needs to meet the high efficiency
remove major barriers in key areas and fuel growth and demand even in partial loads, this can be achieved by
employment. At the same time, the proposals will designing and employing effective control strategies in the
dramatically reduce Europe's dependence on imported oil and
dc-dc converter design and the management of power on bi- converter may require additional work relating to algorithm
directional basis is an essential feature of a dc-dc converter. modifications to achieve better control of both transient and
steady-state performance. These modifications include the use
of a control dead zone, an averaging digital filter to
counteract switching noise, and the use of two sets of gains
one for the transient period and the second for steady-state
performance. However, the application of PID-type
controllers for buck converters is further complicated in a
digital implementation, because digital implementation
requires additional modifications for PID control such as,
dead-zone, additional filtering and gain scheduling [1,2, 6].

The control strategy for the proposed design of the converter

is primarily based on a voltage-mode controlled Pulse Width
Modulation (PWM) technique with a proportional-integral-
derivative (PID) where the control signal is fed from the
output voltage of the converter. However the current-rejected
PWM as the Current-mode control technique can be used to
enhance the control response of the system, bearing in mind
the difficulty to analyse the latter technique because of its
multi-loop architecture.

However, through the research and initial experimental work

conducted with MATLAB/SIMULINK it has been found that
a classical PID control is not always capable of maintaining
the nominal operating point in the presence of disturbances
and modelling errors. Therefore, it is essential that the PID
controller meets specific control requirements, especially
when disturbance rejection and transient response time
requirements are concerned. In addition, due to the non-
linearity of DC-DC converters the classic PID controller does
not always fulfil the control requirements and the PID
Figure 1: Simulink model of the complete dc-dc converter-based controller would need to be modified to a PI controller instead
centralised system architecture [9]. As a result of these findings, there is a need for
developing more intelligent and robust control strategies.
These can be microcontroller or DSP based technologies, as
these two approaches have been used to implement
advanced/improved control algorithms such as nonlinear PID,
Fuzzy Logic, and Feed-forward Control [4, 7].

3 State space modelling of the dc-dc converter

The DC-DC converter is deemed as a variable structure
system, i.e. the circuit topology changes in accordance with
the switching action of the semiconductor device. The state
model is essential as it describes the dynamic behaviour of the
circuit during each portion of the switching period. In the case
of the buck converter, during the ON period the circuit
Figure 2: Dc-dc system architecture contains the source voltage Vs, and during the OFF period the
circuit is source free. Figure 3, shows the circuit topologies
2 Dd-dc converter control techniques for the buck converter in both ON and OFF states.
A key challenge in controlling DC-DC converters is the act of By applying KVL and KCL to figure 3a, a differential
compensating for the converter’s nonlinear damped dynamics equation that describes the dynamic behaviour of the buck
which are a function of load parameters. In recent years converter during the ON-period is obtained:
advances in signal processing technology have spurred
research in new control techniques to improve converter
control. The PID controller is not a rational function and
di
VS L v (1)
research has shown that PID-based control of the power dt
 This mathematical model suggests that the buck converter is a
single-input linear system with constant coefficients.
However, when the switching device is incorporated the
converter exhibits a highly non-linear characteristic. A
simplified model with switching action taken into account is
as follows:
ª q' (t ) º ª q ' (t ) º
ªi º  ªi º «
d « » « 0 L » « »  « L » Vs
»
« q ' (t ) q' (t ) »
dt « » «  »
« » « »
¬v ¼ ¬ c RC ¼ ¬v ¼ « 0 »
¬ ¼ (9)
Where q’ (t) is the switching function.
Figure 3: Topologies of the buck converter
The above analysis was conducted for a single stage,
Where Vs: is the voltage supplied by the alternator, v: is the
however, the proposed converter has six interleaved cells in
output voltage, i: is the circuit current and iC: is the capacitor
order to meet the current requirements for 42V power-net.
current, and R: is the Load.
Figure 4 shows the proposed multiphase DC-to-DC converter.
And the supply current is given by;

dv v 4 Controller Design
i C  (2)
dt R Recent research proposed the use of PI controllers to control
During the OFF-period as illustrated in figure 3b, the inductor soft switched power converters. The classic PI controllers for
is disconnected from the source and short circuited to the power converters are usually designed based on two main
ground, thus equation 1.1 becomes, criteria, maximum load and minimum line condition.
However, as mentioned earlier, due to the non-linearity
di
0 L  v, (3) associated with power electronic converters, they are exposed
dt to variations in operating points over a wide range. It is
While equation 2 remains unchanged during that period, and essential that the PI controller is designed to provide optimal
equations 1 and 3 are elements of a binary control function performance at all operating conditions of the power
defined as follows, converter.

u ^ 1, 0 d t d ton

0 , ton d t d T
(4)
M o s fe t
P h a s e C u rr e n t
L
T o ta l C u r re n t

42V
With equation 4, equations 1 and 3 are combined to form: D io d e C R

di
uVS L v (5) L
dt
By solving equations 2 and 5, the dynamic equations of the
L
buck converter over the entire switching period are obtained:

di v VS P ha se
  u. (6) S h ift
L Vo
dt L L
dv 1 1
i v. (7)
dt C RC L

With equations 6 and 7, the state space model of the buck

converter for the entire switching periods is as follows: L

ªi º ª 1 º ªi º ªVs º
« 0  « »
d « » L » « »« L » (8) M u lti -P h a s e C o n tro lle r
«1 1 »« » « »
dt « » «  » v
¬v ¼ ¬c RC ¼ ¬ ¼ «¬ 0 »¼
Figure 4: The 42V/14V multiphase DC-DC converter
To achieve the required performance the PI Controller is
developed to control the duty cycle of the PWM within the
power converter. The PI controller is designed based on the
state space model of the buck converter in the previous
section [5, 11].

The PI controller was designed using the system matrices

obtained above, the model in figure 5, shows the developed PI
controller for the PWM with switching frequency of 100 kHz,
which represents the frequency of the saw tooth signal in the
model. This controller was developed as a current-mode
PWM controller to eliminate the ringing and oscillation in the
voltage loop and produce a better transient response.

Figure 6: Switching function and PWM control signal

Figure 5: Simulink model of PID (PI) controller for the PWM

switching system

The controller can be implemented using advanced

developments such as nonlinear PI, Fuzzy Logic, and Feed-
forward Control.
Figure 7: Time delay phase shift in Simulink
The PI model developed compares the PWM signal to the
constant saw tooth signal Vpwm, which is 1V in this case.
The frequency of the saw tooth signal is the switching
5 Simulation results
frequency of the converter which is 100 kHz. The output of
the PWM controller is the switching control signal, as shown In order to have an effective control strategy for the buck DC-
in figure 6 which represents a sequence of pulses that drive DC converter a more complex model of the DC-DC converter
the six MOSFETs in the converter. involving the switching element has to be developed. In order
to obtain a structure that is more relevant to real life
As the proposed converter is multi-levelled and has six applications, the control signal is fed to the converter via the
MOSFETs, it is essential to introduce time delay in the PWM, which in turn generates the switching function q’(t).
control signal in order to drive the six active MOSFETs of the The results from initial Matlab model developed with the
converter. By using delay in the Simulink model six phase switching signal q’ (t) and with no feedback control are
control signal waveforms are produced respectively as shown shown in figure 8.
in figure 7.
 Figure 9: Voltage mode controlled PWM response

Figure 8: Inductor current and output voltage of the DC-DC The other commonly used control approach is current-mode
converter with no feedback loop. control (CMC). This method regulates the output current and,
with infinite loop gain, the output is considered to be high. In
From the signals, figure 8, the output voltage has dropped CMC, the current loop is incorporated with a slower voltage
significantly with time, causing the inductor current to loop, and this arrangement can cause difficulty in analysis and
increase. In order to meet the power requirements of the design. However the system response is smoother and there is
system, 14V at around 72A, a carefully designed feedback neither ringing nor oscillation in the system response. Figure
loop is needed. A proper assessment of the open loop gain of 10 shows the output voltage and current of DC-DC converter
the system is essential to help modify the final loop gain with CMC employed.
response for a more stable system operation.
Both responses show that the system finally settles to 14V at
A PI loop to control the duty cycle of the PWM to drive the around 71A producing around 1kW of power at the output of
DC-DC converter the system to improve the response is the system. The input voltage is 42V, with loads resistance set
implemented for both the voltage and the current modes. WR  ȍ +RZHYHU WKHVH FDQ EH HDVLO\ DGMXVWHG LQ Whe
Simulink model to suite the required specifications. The
These two strategies have been investigated and tested and developed model can also be adjusted to consider parasitic
simulation results have been analysed and compared. Figure 9 components in order to create a model that is close to real life
shows the DC-DC converter response with voltage mode applications.
controlled PWM.

This control mode is classified as a voltage-mode controller

(VMC) because the feedback loop regulates the output
voltage of the DC-DC converter. For analysis purposes it is
assumed that if the loop gain is infinite, the output impedance
for an ideal voltage source is considered to be zero.

It is clear from the simulation results that both the voltage and
the current fluctuate significantly before settling to the
required level. In voltage mode control the inductor current is
not monitored nor controlled. This becomes an issue because
the inductor, in conjunction with the output filter capacitor,
forms a resonant tank that can ring and cause oscillations as
the simulation results have identified.

Figure 10: Current mode controlled PWM response

Previous work on this project used the voltage mode control, Digital Controllers for Buck and Boost Converters, in
it was noted that there are significant variations in load Proc. 17th Annu. IEEE Appl. Power Electron. Conf.
current and output voltage depending on driving conditions. Expo., Dallas, TX, Feb. 2002, pp. 381-388.
For instance, a full load operation is rarely present for a
significant period of time, and also, most loads run in stand- 5. M. Lazar, R. De Keyser: Non-linear predictive control of
by mode. However, the effects of these variations are a DC-to-DC converter. SPEEDAM'04 Symposium on
significantly reduced due to the introduction of the current Power Electronics, Electrical Drives, Automation &
mode control for the PWM. Motion, 16-18 June, Capri, Italy; 5p paper nr. A206.

6 Conclusions 6. K. Rajashekara, “42V architectures for automobiles,” in

Proc. IEEE Elect. Manufact. Coil Winding Expo,
The work in this paper resulted in successful Simulink model Indianapolis, IN, pp.431–434.Sep, (2003).
development simulation of the 5kW/42V DC/DC converter
for automotive applications. Also, investigation into the 7. R.P. Severns, G.E. Bloom: Modern DC-to-DC Switch
potential control strategies was conducted and results were Mode Power Converter Circuits. New York: Van
produced to validate the developed designs. Nostrand Reinhold, 1985.

Significant work on Simulink and Matlab was produced to 8. M. Shrud, A. Bousbaine, R.Thorn, “ 42V Power System
validate the mathematical models of systems and control Architecture Development” The 3rd IET Conference on
strategies. The current control mode was developed and found Automotive Electronics, 28-29 June 2007, The
to be more efficient than the voltage control mode used University of Warwick, Coventry, UK.
previously in this field. The DC/DC converter was
successfully developed to convert 5kW/42V produced by the 9. O.J.M. Smith: Posicast Control of Damped Oscillatory
alternator system to 1kW/14V at the output to power lower Systems, Proc. IRE, Vol. 45, No. 9, 1957, pp. 1249-1255.
power components.
10. T. Teratani, K. Kurarnoki, H. Nakao, T. Tachibana, K.
A 42 V system would offer far greater power availability to Yagi, and S. Abae, Development of Toyota mild hybrid
the vehicle electrical system and enable many of the systems (THS-M) with 42 V power net, Records of
traditional mechanical and hydraulic systems to be IEEE–IEMDC–2003, Madison, WI, 2003, pp. 3–10.
electrically driven. Moreover, the current drawn from the
battery will be lower and the possibility to implement 11. Zeller, J.; Zhu, M.; Stimac, T.; Gao, Z.: Nonlinear Digital
additional safety and comfort features. In addition, nearly all Control Implementation for a DC-to-DC Power
accessory systems present in today’s vehicles are candidates Converter. Proc. of 36th Intersociety Energy.
for electrical conversion. These include a variety of pumps,
blowers, and actuators that can benefit from the introduction
of variable-speed motor drives.

7 References
1. J.M.Altes, E.G. Dolcet, and B.P. Solorzano. “Analysis of
the Most Appropriate Electrical Architecture and
Communication Bus for the New Dual Voltage 14/42v
System”, IECON-2002: Proceedings of the 2002 28th
Annual Conference of the IEEE Industrial Electronics
Society, Vol.1-4 ,pp. 1687-1692, (2002).

2. Ali Emadi, Sheldon S. Williamson, and Alireza

Khaligh," Power Electronics Intensive Solutions for
Advanced Electric, Hybrid Electric, and Fuel Cell
Vehicular Power Systems" IEEE Transactions on Power
Electronics.VOl.21, NO.3, May, (2006).

3. Department for Transport (2004): Powering Future

Vehicles. The Government Strategy, Second Annual
Report. London: Department for Transport.

4. L. Guo , J.Y. Hung, R.M. Nelms: PID Controller

Modifications to Improve Steady-State Performance of