IEEE Electrification Magazine - March 2017 - 30

include mass, volume, mechanical mounting features,
coolant temperature range, and flow rate.
Environmental requirements for the inverter are determined based on the mounting location within the vehicle.
Environmental stress factors, such as vibration, temperature cycling, and thermal shock, all play a tremendous role
in the design and selection of components for the EV
inverter. For example, the mechanical design must consider the vibration source excitation seen at the inverter.
There is often a requirement placed on internal mechanical resonant frequencies to be greater than some minimum value (for example, 400 Hz) to avoid exciting the
resonances during a normal drive cycle. Structural finite
element analysis can identify any unwanted resonances
early in the design phase and help avert problems later
on. Once physical hardware is available, test sweeps on a
vibration fixture can be used to empirically identify key
resonant frequencies. Resonance can result in mechanical
fatigue at high-stress locations and cause premature failure if not properly damped.
EVs must operate in a diverse range of geographical
locations and climates. Inverter operating temperature is
influenced by both the liquid coolant and the surrounding
ambient air temperature. Almost all high-power EV inverters use liquid cooling (typically a water-ethylene glycol
mixture) to provide the highest density package possible.
The maximum coolant inlet temperature and flow rate
are critical parameters for the power semiconductor cooling, and the inverter current rating will largely be driven
by the heat removal capability of the liquid cooing system
and associated thermal impedances.

hVdc Bus Filtering
Starting at the input to the inverter, we can first examine
the HVdc interface. There are multiple constraints that
drive the design of the HVdc filtering employed in EV
inverters. First, the voltage overshoot on power semiconductor switches during turn-off transition (Figure 3) must
be limited to stay below the semiconductor device rating.

Ch1/Ch2: Vge, 5 V/div
Ch3: Vce, 100 V/div
Ch4: Ic, 200 A/div

Figure 3. The switching waveforms for an insulated-gate bipolar transistor (IGBT)-based inverter at 400 Vdc and 1,000 Apk. Vge: gate to emitter
voltage; Vce: collector to emitter voltage; Ic: collector current; div: division.

I E E E E l e c t r i f i c ati o n M agaz ine / march 2017

This is achieved by ensuring a low impedance path across
the dc link, such as placing a high-quality bypass capacitor close to the power semiconductor switches in a lowinductance loop (capacitor Cx in Figure 2). Second, a
source of local energy storage is required, as the battery
may be located several feet from the inverter and the
cable inductance from battery pack to inverter can be as
large as a few microhenries. Third, we would like to minimize large ripple currents generated by the inverter pulsewidth modulation (PWM) switching from flowing back
toward the dc bus and into the battery or other components (Anwar et al. 2010). Finally, we want to provide highfrequency common mode (CM) and differential mode
filtering to minimize electromagnetic interference (EMI)
generated by the inverter.
In industrial style inverters, it is common to use large
electrolytic capacitors for HVdc energy storage. Due to
packaging constraints, these may be located several inches from the power semiconductor devices. In such an
inverter, it may be necessary to place discrete snubber
capacitors close to the switching power device and possibly even right on the power module terminals. In comparison, most EV inverters use large polypropylene film
capacitors to provide local energy storage and bypass.
These are connected to the power devices using very-lowinductance bus bar structures. In many cases, the need for
discrete snubber capacitors can be eliminated, which
saves cost and volume. This requires careful attention to
the internal design of the bulk capacitor and external bus
bar to achieve sufficiently low stray inductance connection to the power device.
The basic specifications for the bulk capacitor must
include working voltage, root-mean-square (rms) current,
duty cycle, lifetime, operating temperature, and parasitic
inductance and resistance. The current requirements are
often specified for several different use cases, representing
various driving events. For example, 0- to 60-mi/h maximum acceleration (high current for very short duration) or
highway cruising up a long steep grade (moderate current
for a long time) are two common events. Each case is
defined by current amplitude, duration, as well as the
cumulative number of times these events are expected to
occur over the life of the capacitor. These can be derived
from the inverter mission profile. Due to the large physical
size and thermal time constant of the bulk capacitor, it is
often a long duration event, such as 10-20 min, that proves
to be the most challenging specification to meet.
The capacitor construction most commonly employs
polypropylene film, and the volume of the capacitor is
roughly proportional to the square of the film thickness.
To minimize component size, the trend has been to push
for continually thinner films. Today's technology uses less
than 3-µm film thickness for capacitors in the 400-Vdc
range. Special segmented metallization patterns are also
employed to extend the high-temperature lifetime of the
capacitor and provide a more gradual decrease in

Table of Contents for the Digital Edition of IEEE Electrification Magazine - March 2017