62 Focus | Inverters March/April 2024 | E-Mobility Engineering introduce a lot more switching noise,” he says. “At a certain point it starts to balance out how much space you can save, because you then need to introduce larger filter elements, so there’s a certain sweet spot that technologically we can’t get through at this point.” The packaging of the power semiconductor devices also has a strong impact on power density, says our expert in electric machines, power electronics and control systems. “A modular design, optimised for reuse using individual halfbridges won’t be as dense as another one optimised for a specific application using a six-pack configuration. Architectural decisions should be made considering this trade-off,” he explains. The DC link capacitor can amount to up to 30% of an inverter’s volume, he adds. “Its volume can be minimised by accepting a higher-voltage ripple or using switching strategies to reduce it. Size reduction can also be achieved by using thinner insulation materials or new types of material.” While power density is strongly affected by cooling performance, extra cooling can indirectly reduce the efficiency of an inverter based on SiC metal-oxide semiconductor field-effect transistors (MOSFETs) if the designer chooses to push more current through them. This is because they present resistance and their losses rise with the square of the current, the traction inverter developer expert points out. In the company’s latest inverter family, the solution was to run more MOSFETs in parallel to meet the current requirement while minimising the resistance between the drain and source terminals with the switch ‘on’, and using a low-cost, lowperformance thermal junction to a simple heat sink. A question of topology Inverter designers can draw on different topologies to meet the needs of their applications. The topology is the specific arrangement of the components and circuitry, which defines how the switches, capacitors and inductors are connected, for example, and how they interact to achieve the desired DC to AC (and vice versa) conversion. Most of the inverters used in e-mobility applications are of the voltage-source inverter (VSI) type, which are designed to provide a constant voltage output and adjust the output current to meet load requirements. Most of these are of the two-level type that switch between two voltage levels, which typically correspond to the positive and negative values of the system’s DC bus voltage. A third level of 0V is achieved in the ‘off’ portion of the duty cycle, which allows the motor current to freewheel through the power devices, rather than the DC link capacitor, reducing the ripple current on that component. However, multi-level inverters (MLIs) are attracting interest for high-power applications. MLIs switch between multiple voltage levels to approximate a sinusoidal waveform more closely, and they can achieve higher voltages with lower harmonic distortion and reduced stress on components at the cost of increasing the complexity of the inverter hardware and controls. There are different topologies of multilevel inverters, including neutralpoint clamped (NPC) and three-level neutral-point clamped (T-NPC), for example. Their benefits vary according to the duty cycle, says the specialist in electric machines, power electronics and controls. The basic NPC topology provides three voltage levels. It has a neutral point that is clamped to the midpoint of the DC bus voltage, creating three voltage levels. T-NPC extends the NPC topology to provide additional voltage levels; adding an extra switching leg to increase the total number of voltage levels to five. As might be expected, MLIs with T-NPC topology can create a smoother output waveform with reduced harmonic distortion at the cost of greater complexity, particularly when it comes to control software. Despite their benefits, multi-level inverters essentially present the same performance limitations as a conventional two-level, three-phase system in providing a wide range of speed and torque, says our traction inverter developer’s expert. The company has tackled this in its latest inverters with a new use for the established technique of switching the motor-winding connection scheme between series and parallel in real time. “You’re changing the effective number of turns on the motor’s stator coils. In series mode, you have twice as many turns in parallel, so at low speed you can effortlessly create large amounts of torque. Then, as the speed increases, the machine is switched to parallel, so the back EMF is cut in half. Torque is also cut in half, but now the power delivery at speed is achieved relatively easily and efficiently since field weakening is not required until much higher speeds are reached, if at all. “We gain a lot of efficiency on the motor side because we don’t have to push the field-weakening D-axis current into the motor to enable it to produce torque at high rpm. If you take a machine from 90% efficiency to 95% at high speed, which is where most of the power is consumed, it has a much more significant impact on the efficiency of a vehicle.” Dana’s HD motor/inverter system has a nine-phase HV inverter with variable switching frequency, Reflex technology gate drivers and an EMI filter (Image courtesy of Dana TM4)
RkJQdWJsaXNoZXIy MjI2Mzk4